Durability and Reliability of Polymers and Other Materials in Photovoltaic Modules (Plastics Design Library) 0128115459, 9780128115459

Citation preview

Plastics Design Library Series

DURABILITY AND RELIABILITY OF POLYMERS AND OTHER MATERIALS IN PHOTOVOLTAIC MODULES

Edited by Hsinjin Edwin Yang Pioneer Scientific Solutions, LLC, Illinois, United States

Roger H. French Case Western Reserve University, Cleveland, OH, United States

Laura S. Bruckman Case Western Reserve University, Cleveland, OH, United States

PLASTICS DESIGN LIBRARY (PDL) PDL HANDBOOK SERIES Series Editor: Sina Ebnesajjad, PhD ([email protected]) President, FluoroConsultants Group, LLC, Chadds Ford, PA, United States www.FluoroConsultants.com The PDL Handbook Series is aimed at a wide range of engineers and other professionals working in the plastics industry, and related sectors using plastics and adhesives. PDL is a series of data books, reference works and practical guides covering plastics engineering, applications, processing, and manufacturing, and applied aspects of polymer science, elastomers, and adhesives. Recent titles in the series Polymeric Foams Structure-Property-Performance, Obi (ISBN: 9781455777556) Technology and Applications of Polymers Derived From Biomass, Ashter (ISBN: 9780323511155) Fluoropolymer Applications in the Chemical Processing Industries, 2e, Ebnesajjad & Khaladkar (ISBN: 9780323447164) Reactive Polymers, 3e, Fink (ISBN: 9780128145098) Service Life Prediction of Polymers and Plastics Exposed to Outdoor Weathering, White, White & Pickett (ISBN: 9780323497763) Polylactide Foams, Nofar & Park (ISBN: 9780128139912) Designing Successful Products With Plastics, Maclean-Blevins (ISBN: 9780323445016) Waste Management of Marine Plastics Debris, Niaounakis (ISBN: 9780323443548) Film Properties of Plastics and Elastomers, 4e, McKeen (ISBN: 9780128132920) Anticorrosive Rubber Lining, Chandrasekaran (ISBN: 9780323443715) Shape-Memory Polymer Device Design Safranski & Griffis (ISBN: 9780323377973) A Guide to the Manufacture, Performance, and Potential of Plastics in Agriculture, Orzolek (ISBN: 9780081021705) Plastics in Medical Devices for Cardiovascular Applications, Padsalgikar (ISBN: 9780323358859) Industrial Applications of Renewable Plastics, Biron (ISBN: 9780323480659) Permeability Properties of Plastics and Elastomers, 4e, McKeen (ISBN: 9780323508599) Expanded PTFE Applications Handbook, Ebnesajjad (ISBN: 9781437778557) Applied Plastics Engineering Handbook, 2e, Kutz (ISBN: 9780323390408) Modification of Polymer Properties, Jasso-Gastinel & Kenny (ISBN: 9780323443531) The Science and Technology of Flexible Packaging, Morris (ISBN: 9780323242738) Stretch Blow Molding, 3e, Brandau (ISBN: 9780323461771) Chemical Resistance of Engineering Thermoplastics, Baur, Ruhrberg & Woishnis (ISBN: 9780323473576) Chemical Resistance of Commodity Thermoplastics, Baur, Ruhrberg & Woishnis (ISBN: 9780323473583) Color Trends and Selection for Product Design, Becker (ISBN: 9780323393959) Fluoroelastomers Handbook, 2e, Drobny (ISBN: 9780323394802) Introduction to Bioplastics Engineering, Ashter (ISBN: 9780323393966) Multilayer Flexible Packaging, 2e, Wagner, Jr. (ISBN: 9780323371001) Fatigue and Tribological Properties of Plastics and Elastomers, 3e, McKeen (ISBN: 9780323442015) Emerging Trends in Medical Plastic Engineering and Manufacturing, Scho¨nberger & Hoffstetter (ISBN: 9780323370233) Manufacturing and Novel Applications of Multilayer Polymer Films, Langhe & Ponting (ISBN: 9780323371254) PEEK Biomaterials Handbook, 2e, Kurtz (ISBN: 9780128125243) Fluoropolymer Additives, 2e, Ebnesajjad (ISBN: 9780128137840) The Effect of UV Light and Weather on Plastics and Elastomers, 4e, McKeen (ISBN: 9780128164570) To submit a new book proposal for the series, or place an order, please contact Edward Payne, Acquisitions Editor at [email protected]

William Andrew is an imprint of Elsevier The Boulevard, Langford Lane, Kidlington, Oxford, OX5 1GB, United Kingdom 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States Copyright Ó 2019 Elsevier Inc. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-12-811545-9 For information on all William Andrew publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Matthew Deans Acquisition Editor: Edward Payne Editorial Project Manager: Lindsay Lawrence Production Project Manager: Bharatwaj Varatharajan Cover Designer: Greg Harris Typeset by TNQ Technologies

Contributors Jennifer L. Braid Case Western Reserve University, Case School of Engineering, SDLE Research Center, Materials Science and Engineering, Cleveland, OH, United States Laura S. Bruckman Case Western Reserve University, Case School of Engineering, SDLE Research Center, Materials Science and Engineering, Cleveland, OH, United States David M. Burns Chairman, ASTM Committee G03.08 on Service Life Prediction, Woodbury, Minnesota, United States Kristopher O. Davis Department of Materials Science and Engineering, University of Central Florida, Orlando, Florida, United states Justin S. Fada Space and Science Technology Systems Branch, John H. Glenn Research Center, NASA, Cleveland, Ohio, United States; Department of Mechanical and Aerospace Engineering, Case Western Reserve University, Cleveland, Ohio, United States Andrew Fairbrother Engineering Laboratory, National Institute of Standards and Technology (NIST), Gaithersburg, MD, United States Sean Fowler Q-Lab Corporation, Westlake, Ohio, United States Roger H. French Case Western Reserve University, Case School of Engineering, SDLE Research Center, Materials Science and Engineering, Cleveland, OH, United States Abdulkerim Gok Gebze Technical University, School of Engineering, Materials Science and Engineering, Gebze, Kocaeli, Turkey

Devin A. Gordon Case Western Reserve University, Case School of Engineering, Macromolecular Science and Engineering, Cleveland OH, United States; Case Western Reserve University, Case School of Engineering, SDLE Research Center, Cleveland OH, United States Xiaohong Gu Engineering Laboratory, National Institute of Standards and Technology (NIST), Gaithersburg, MD, United States Yang Hu GE Renewable Energy, San Ramon, CA, United States Long Jiang NDSU, Fargo, ND, United States Sumanth Varma Lokanath First Solar Inc, Mesa Arizona, United States Ina T. Martin The Materials for Opto/Electronics Research and Education (MORE) Center, Case Western Reserve University, Cleveland, Ohio, United States Antonia Omazic Polymer Competence Center Leoben, Leoben, Austria Gernot Oreski Polymer Competence Center Leoben, Leoben, Austria Bettina Ottersbo¨ck Polymer Competence Center Leoben, Leoben, Austria Timothy J. Peshek Photovoltaics and Electrochemical Systems Branch, John H. Glenn Research Center, NASA, Cleveland, Ohio, United States Nancy Phillips E. I. du Pont de Nemours and Company, Wilmington, DE, United States

xiii

xiv

Eric John Schindelholz Sandia National Laboratories, Albuquerque, New Mexico, United States Eric J. Schneller Department of Materials Science and Engineering, University of Central Florida, Orlando, Florida, United states Bryan Skarbek First Solar Inc, Mesa Arizona, United States Menghong Wang Case Western Reserve University, Case School of Engineering, Macromolecular Science and Engineering, Cleveland OH, United States; Case Western Reserve University, Case School of Engineering, SDLE Research Center, Cleveland OH, United States

C ONTRIBUTORS

Kent Whitfield Renewable Energy Technologies, Fremont, United states Hsinjin Edwin Yang Pioneer Scientific Solutions, LLC., Long Grove, IL, United States Shuying Yang Tesla Inc, Fremont, CA, United States Allen F. Zielnik Atlas Material Testing Technology LLC, IL, United States

1 Overview: Power Degradation and Failure of PV Systems1 Jennifer L. Braid and Roger H. French Case Western Reserve University, Case School of Engineering, SDLE Research Center, Materials Science and Engineering, Cleveland, OH, United States

Understanding the value of increased or decreased module reliability at the PV systems level is critical to successfully manage PV module reliability. Reliability directly influences the economic viability of photovoltaics as an energy source not only by controlling the total number and size of revenue payments received from the future sales of electricity, but also by influencing O&M costs, and the cost of money required to build the PV system. After considerations of present-value discounting and escalation of the worth of electricity in future years, a 30-year PV plant, for example, can be worth 25%e30% more than a 20-year-life plant. Based on this economic sensitivity to plant life and the billion dollar cost of a utilityscale PV power plant today, there is a strong incentive to strive for a long life for such systems and a large incentive to allocate substantial funds for improving reliability. R. G. Ross [1]

Chapter Points
The successful development of the crystalline silicon (c-Si) photovoltaic (PV) technology has always been tied to the reliability and lifetime performance of said technology and the packaging technologies of the PV module.
The transition from lab-scale technology development and demonstration to real-world application of photovoltaic power plants spans 1

Our focus here is on photovoltaics based on crystalline silicon cells, as opposed to the many other photovoltaic technologies that have been demonstrated or commercialized at smaller scales.

required technological refinements for reliability and durability, which are often neglected in lab-based research.
Throughout the 40 year history of c-Si PV technology development, these technology scale-up transitions have been associated with some of the largest reliability fiascos for photovoltaics, illustrating the challenging nature of transition from the lab to the real-world.

1.1 Introduction The SunShot initiative supported by the US Department of Energy (DOE) aims to advance and accelerate the development and growth of photovoltaic (PV) energy in terms of efficiency, reliability, manufacturing, and solar electricity costs by identifying priority research directions for PV research and technology development [2]. By setting goals toward many scientific and technological challenges and funding these research needs, SunShot’s 2020 cost target for the levelized cost of electricity (LCOE) of $0.06 per kilowatt-hour (kWh) for utility-scale PV was achieved in 2017 (compared to $0.27/kWh in 2010), and continued efforts are being made to achieve $0.03/kWh by 2030 (Fig. 1.1) [3]. PV power plant development and the associated PV module installations also continue their rapid global expansion to a total of 98 GW installed globally in 2017 [4e6], with continued growth as shown in Fig. 1.2. The total installed capacity at the end of 2017 was approximately 402 GW (29% compound annual growth rate compared to 225 GW total installed capacity in 2015 and 76 GW in 2016) representing more than 2% of the global electricity consumed [7e9]. According to the International Energy Agency’s latest report, “World Energy Outlook 2017”, renewable energy sources, including solar,

Durability and Reliability of Polymers and Other Materials in Photovoltaic Modules. https://doi.org/10.1016/B978-0-12-811545-9.00001-X Copyright © 2019 Jennifer L. Braid & Roger H. French. Published by Elsevier Inc. All rights reserved.

3

4

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 1.1 Levelized cost of electricity (LCOE) for utility, commercial, and residential PV systems, with costs in 2010 and 2017 and SunShot goals for 2020 and 2030. LCOE is quoted in 2017 dollars. Source: Department of Energy, Photovoltaics, 2018. https://energy.gov/eere/sunshot/photovoltaics.

Figure 1.2 Globally installed PV capacity with projection up to the year 2022 (GW). Source: International Renewable Energy Agency, Data and Statistics, IRENA, 2017. http://resourceirena.irena.org/gateway/ dashboard/?topic¼4&subTopic¼16

1: OVERVIEW: P OWER D EGRADATION

AND

FAILURE

OF

are anticipated to replace fossil fuels in the next 25 years while expecting a 30% increase in the global energy demand over the same time period [8,10]. In order to address the future energy supply and compete effectively with fossil fuels and thus diminish carbon emissions, PV energy is of vital importance as a sustainable, renewable, and clean energy alternative. The growing importance of PV, compared to fossil fuels, is confirmed by the market’s rapid increase of PV installations and the increasing contributions of PV to electricity generation. At the same time, coal and nuclear installations have stalled, and their contributions to total electricity production are expected to decrease from now to 2040. Oil demand, while still growing, is doing so at decreasing annual rates [8].

1.2 Photovoltaic Technology Development, Insertion, and Growth The modern Si solar cell was invented at Bell Laboratories in 1954, with an impressive 6% power conversion efficiency [11]. In the early years, solar cells were mostly employed for remote communication stations or small novelty-type applications, like powering transistor radios and glasses-mounted hearing aid battery chargers. Many of these systems performed poorly due to inadequate protection from weather or ambient conditions [12]. Photovoltaics research quickly found the perfect incubator in the space race. Sputnik was launched by the USSR in 1957 and was powered with electrochemical batteries. By then, research Si cells had reached 15% efficiency [13], and the US launched the first solarpowered satellite, the Vanguard I, in 1958. The Vanguard I’s PV array consisted of six 18-cell panels distributed over its spherical shell. Cells were 0.5  2 cm n-type (phosphorous-doped) and operated at approximately 10% efficiency, for a total array power output of 8% by 2020 of total power generation for power producers above 5 GW [65]. Most government programs, however, focused on increasing PV manufacturing rather than promoting installations. The 11th five-year plan (2006e10) stated an objective to “Encourage the production and consumption of renewable energy and increase its share in total primary energy consumption”, but in practice drove economic growth over environmentalism [66]. Chinese PV production increased by 70%/yr from 1997 to 2007 [61], but China also became the world’s largest consumer of electricity and emitter of CO2 in 2007 [64]. China held 35% of the world PV production capacity in 2008, but exported 98% of the modules it manufactured [63]. The status of China as a global competitor in the PV market was still in question, too. Japan had increased production by 50%/yr since 1995 and held half the world production in 2007, with other countries trailing behind at an average growth rate of 38%. While these rates were considerably lower than China’s, its challenges were in securing silicon which was in global shortage, developing a domestic market for PV modules, and stimulating domestic innovation in PV technology [61]. In order to secure its place in the global PV market, China began to focus R&D efforts on cell

PV S YSTEMS

13

materials, specifically crystalline silicon production [64]. Vertically integrated PV companies managing the supply of c-Si, wafering, cell, and module lines became the norm, and pushed the raw material selfsufficiency of Chinese PV manufacturing from 0% in 2006 to 50% in 2010. China produced 50% of the world’s PV modules in 2010 and 60% in 2011. However, the global market lagged behind, with this production doubling global demand in 2012 [65]. Worsening the situation, the US and EU imposed antidumping and countervailing duties to curb imports of low-quality modules. Chinese PV module exports decreased from 22.77B USD in 2011 to 12.78B USD in 2012 [67]. The Chinese government had invested heavily in the PV market (33.7B USD in 2010 alone, with 30B USD in low-cost loans going to the top five Chinese PV manufacturing companies). Additionally, China was subsidizing domestic PV installations, increasing the total installed capacity to 12.92 GW in 2013, but the domestic market was still underdeveloped [67]. By 2012, the top 10 PV manufacturers had a combined 111B RMB in debt [65], and all Chinese PV producers were operating at a loss from 2011 to 2014. This financial crisis, along with the production capacity surplus, led to widespread bankruptcy and restructuring of the domestic PV industry in China in 2012 [64]. The Chinese government, in an effort to prevent collapse of the domestic PV industry, continued to heavily fund PV R&D and provide loans to major manufacturers. Policies were advanced to improve domestic technology for polysilicon and cell production, and close the gap in energy consumption, utilization rates, and quality level of Japan, the US, and the EU. Chinese polysilicon reached 43% of global production in 2014, but its purity was still lower than that of the US or Germany [67]. Meanwhile, with PV module production having increased 1000x from 2000 to 2012 [64], the LCOE of PV electricity had dropped to a more competitive $0.211/ kWh [68]. As the price of modules continued to drop, worldwide installations increased and PV became a more widely recognized and accepted grid-connected energy source. The 12th five-year plan also mandated that nonfossil fuels would comprise 11.4% of all energy consumed in China by 2015, so domestic installations also grew [65]. Many Chinese PV module companies also began to shift development strategies, moving into downstream industry (energy system development/provider) or expanding into international markets [67]. The policy-driven market

14

D URABILITY AND R ELIABILITY

OF

P OLYMERS

(both Chinese and international) continued to fuel the Chinese PV industry, with PV LCOE reaching $0.06/ kWh in 2017 (3 years ahead of the US DOE SunShot 2020 goal).

1.4 PV Module Degradation To this point, we have used the terms “durability” and “reliability” almost interchangeably. However, these terms have very specific meanings in the field of module degradation: durability is related to slow, gradual performance loss, while reliability signifies resistance to sudden failures [17]. Both durability and reliability are important to the lifetime and degradation of a PV module, as well as to the success of the PV industry as a whole. Module failure, whether a result of durability or reliability issues, is generally defined as a 20% decrease in power conversion efficiency [24]. Various studies and module warranties have defined module failure in the range of 25%e50%, particularly for high-efficiency modules, which may still have higher power output in this range than a nondegraded low-efficiency module. In today’s PV market, modules are usually offered with a 10 year limited product warranty and a 20e25 year power performance warranty with an estimated 1% power loss each year (at least 75% power efficiency at the end of 25 years) [69,70]. For multi- or mono-crystalline silicon PV cells, there is often a larger permissible power loss of 1%e3% in the first year after installation. Some manufacturers provide a 25 year performance guarantee with less than 0.7% power loss per year, and there have even been 30 and 35 year performance guarantees in the market [71]. The problem is that the same warranty applies for the same modules deployed in all climatic zones where modules face very different environmental conditions such as extreme hot or cold, humidity, heavy snow loading, or gusty winds. Material selection and module design are therefore key parameters for climate sensitive degradation and there is a strong research need in the PV community to address weather induced reliability issues. Lifetime performance of PV systems requires a better understanding of degradation mechanisms of materials, components, and systems observed in the real world. The encapsulant is widely accepted to be the largest source of module degradation, as it does not enhance performance or reliability versus

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

unencapsulated cells [1,72e74]. Rather, it “often aggravates or creates a number of failure mechanisms, which include cracking, yellowing, delamination, accelerated corrosion, and differential expansion stresses, voltage breakdown, excessive leakage currents, increased soiling, or increased operating temperatures.” [1] Encapsulant browning observed at Carrisa Plains and other sites was considered a disaster at the time, but the actual dominant degradation modes in those modules were related to chemical changes in EVA that attacked the cells, not module appearance or optics. EVA is unstable under high operating temperatures and high insolation, which cause deacetylation of vinyl acetate to release acetic acid and discoloring chromophore polyenes. Acetic acid and acetaldehyde produced in this reaction both catalyzed further EVA degradation and corroded the cell metalization and interconnects. This behavior was not reported for PV sites in Eastern/Central US or Western Europe, only for hot/ dry climatic zones, illustrating the inherent problem in certifying and warrantying a module for global use. The phenomenon of EVA browning was already known before observation in fielded PV modules, as it was demonstrated on films in accelerated exposure in 1979. However, most manufacturers were unaware of the problem, and misinterpretations on thermal aging and reaction rates of EVA led to inaccurate extrapolation of lifetime results [17]. The inherent problems with EVA have led to calls for a replacement polymer encapsulant, but cost concerns have continued to dominate and EVA remains the most widely used encapsulant material in c-Si modules. Ross [1] identified the electrical circuit to be the second most problem-prone module component, including cell metalization, interconnects, module junction box, bypass diodes, and module interconnects. Module improvements developed during the JPL-FSA project such as redundant cell interconnects, bypass diodes, and standardized module interconnects greatly improved the reliability of the electrical circuit. However, durability concerns still exist regarding hot spotting, overheating, and leakage currents. The low-cost materials developed during and since the JPL-FSA program have enabled the widespread implementation of photovoltaics for commercial energy production we see today, but these materials also dictate the dominant degradation mechanisms in PV modules. For the cost of solar energy to continue to fall, we must consider the lifetime performance of

1: OVERVIEW: P OWER D EGRADATION

AND

FAILURE

OF

these materials and the effects of their interactions on module durability [75,76]. Robust adoption of PV systems has relied on significant technology developments, not only in terms of increased cell or module efficiency and decreased manufacturing and electricity costs [40,77,78], but also in reliability and durability for long service lifetimes and guaranteed lifetime performance.

1.5 Moving Towards Terawatt PV Energy The dramatic progress of commercial crystalline silicon photovoltaics since their development in the 1970s has included both successes and dramatic shortcomings such as those of the Carissa Plains plant. It is therefore important to assess the state of c-Si PVat this point, what challenges lie ahead, and what the future of photovoltaic electricity generation and the PV industry and value chain will look like. It is essential that we continue to improve not only the solar energy conversion efficiency of PV modules, but also their reliability and lifetimes, while taking advantage of the reduction in the cost of PV modules and PV power plants as the industry scales to larger annual production volumes, and the economies of scale continue reducing the levelized cost of electricity [2]. This means that all scientists, technologists, operators, and financiers must learn from both the successes and failures of photovoltaics that have transpired to date. It is also essential to have a sense of how large the PV industry and PV electricity generation can become globally (i.e. market opportunity), since this is an essential driver for industry growth and market penetration of photovoltaics. For example, there were 402 GW of PV power plants operating globally as of the end of 2017. Is this expected to grow to 1 or 10 TW of installed PV power plant capacity, or is PV electricity too expensive to compete, and 402 GW will shrink in the forthcoming years? Is this the scale we should expect for PV generation in the future (i.e. are the current PV materials, cell, module, and power plant companies now established and mature)? Or is more growth expected, and to what scale do we expect photovoltaics to grow, and how are they competing against out-of-kind electricity generation technology (e.g. coal, nuclear, oil and gas, hydroelectric, and wind generation)? To answer these questions, the research and studies done by the International Energy Agency, as reported in their annual World Energy Outlook

PV S YSTEMS

15

reports [8,10,49], and by the US Energy Information Agency in their annual Energy Outlook for the United States [79], are very useful guideposts that incorporate technological progress [48], societal needs, and the impacts of global and national progress, strife, and conflicts. The IEA World Energy Outlook uses a scenario-based approach to develop projections for global energy demand, supply, and energy sector transformation. In WEO 2018, they have three scenarios, the current policies scenario (CPS), the new policies scenario (NPS), and the sustainable development scenario (SDS). The current policies scenario is based solely on currently enacted laws and regulations and can be considered as the baseline scenario. The new policies scenario also incorporates a number of current policy frameworks that appear likely to be enacted along with current estimates of technological progress. In 2015, the United Nations General Assembly adopted by resolution the 2030 Agenda for Sustainable Development [80], which is a set of 17 sustainable development goals (SDG) that serve as a “blueprint to achieve a better and more sustainable future for all”. A number of these goals are closely coupled to energy, including SDG7: affordable and clean energy, SDG3: good health and well-being, SDG13: climate action, and SDG6: clean water and sanitation. Other goals are coupled with economic development such as SG9: industry, innovation, and infrastructure, SDG11: sustainable cities and communities, and SDG12: responsible production and consumption. Due to the critical role of energy, IEA has developed the sustainable development scenario in 2017 in the World Energy Outlook, which is aligned with “the Paris Agreement, providing universal access to modern energy by 2030 and reducing dramatically premature deaths from energy related air pollution”. Here we will use the SD scenario results from WEO 2018 to provide a vision of where photovoltaics may be heading as the world heads to 2030. For those interested in more information in the alternative scenarios, we direct you to the IEA WEO reports and the UN’s SDG reports and to the World Energy Model [81]. Using these IEA scenarios and the IEA World Energy Model, we summarize in Table 1.2 the scale of the global installed power generation capacity of solar photovoltaics from the 2017 value through 2040. One thing to note is that these IEA WEO projections show linear growth in module production capacity, whereas historically the growth has been

16

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Table 1.2 Total Installed Power Generation Capacity, in Terawatts, for Solar PV Under Different IEA Scenarios

Year

New Policies Scenario

Sustainable Development Scenario

2017

0.4

0.4

2025

1.109

1.589

2030

1.589

2.346

2035

2.033

3.300

2040

2.540

4.240

exponential [82]. Also for 2017, the PV module global production capacity was estimated to be on the order of 80e85 GW of PV modules annually [83], and this was updated in 2018 to have actually been 94e100 GW in 2017 and 105e115 GW of PV cells and modules produced in 2018 [84]. Therefore the new policies scenario, which has a linear growth rate of 92.8 GW of PV modules produced per year, appears to be conservative, not requiring much, or any, growth of the current PV module production capacity to achieve 1.1 TW of installed capacity in 2025 and 2.54 TW of PV power generation in 2040. For the Sustainable Development Scenario, the linear growth rate of PV power generation would require 168 GW of PV module manufacturing capacity globally. And this scenario leads to 1.6 TW installed in 2025 and 4.24 TW of installed PV in 2040. One critique of the IEA WEO reports is that they have been overly conservative in regards to PV module production, and that their forward projections are of linear growth, even if the historical growth observed has been closer to exponential growth [82]. In addition, if the global efforts to meet the sustainable development goals and climate change challenges advance, then it is relatively straightforward to continue growing the size of the PV module manufacturing industry, and further reducing the costs of PV modules and PV generated electricity [85]. These perspectives suggest that even 4.2 Terawatts of global PV power generation in 2040 maybe a conservative estimate. These projections underline how controlling PV module degradation, improving PV module reliability, and extending PV power plant lifetimes are critically important to industry, society, and the world.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Acknowledgments The authors would like to acknowledge the important role played by the funding of the US Department of Energy’s Office of Energy Efficiency and Renewable Energy (EERE) under Solar Energy Technologies Office (SETO) Agreement Numbers DE-EE-0008550 and DE-EE-0008172, and previous agreements DE-EE-0007140, DE-EE0007143, and DE-EE-0004946, Subaward Agreement no. 60220829-51077-T, which have enabled us to learn and develop PV module degradation science. The SDLE Research Center was established through funding by the Ohio Third Frontier, Wright Project Program Award tech 12-004. We also acknowledge the research performed by graduate and undergraduate students, post-doctoral researchers, and our academic, national lab, and industrial collaborators, without which we would not have been able to advance the field of PV degradation and reliability.

References [1] R.G. Ross, PV reliability development lessons from JPL’s Flat Plate solar array project, IEEE Journal of Photovoltaics 4 (1) (2014) 291e298, https://doi.org/10.1109/JPHOTOV.2013.2281102. http://ieeexplore.ieee.org/xpl/articleDetails.jsp? arnumber¼6616584. [2] R. Jones-Albertus, D. Feldman, R. Fu, K. Horowitz, M. Woodhouse, Technology advances needed for photovoltaics to achieve widespread grid price parity: widespread grid price parity for photovoltaics, Progress in Photovoltaics: Research and Applications 24 (9) (2016) 1272e1283, https://doi.org/10.1002/pip.2755. http://doi.wiley.com/10.1002/pip.2755. [3] Department of Energy, Photovoltaics, Department of Energy, 2018. https://energy.gov/eere/ sunshot/photovoltaics. [4] International Renewable Energy Agency, Data and Statistics, IRENA, 2017. http://resourceirena. irena.org/gateway/dashboard/?topic¼4&sub Topic¼16. [5] A. Zervos, Renewables 2017 Global Status Report. Renewable Energy Policy Network for the 21st Century, 2017, ISBN 978-3-9818107-6-9. http://www.ren21.net/gsr-2017/.

1: OVERVIEW: P OWER D EGRADATION

AND

FAILURE

OF

[6] Photovoltaics Report, Fraunhofer ISE, 2017. http://www.ise.fraunhofer.de/en/publications/ studies/photovoltaics-report.html. [7] International Energy Agency, A Snapshot of Global PV 1992e2016, 2017. Tech. rep. http:// iea-pvps.org/fileadmin/dam/public/report/statistics/ IEA-PVPS_-_A_Snapshot_of_Global_PV_-_ 1992-2016__1_.pdf. [8] International Energy Agency, World Energy Outlook 2017, Organization for Economic Co-operation and Development (OECD)/IEA, 2017. https://www.iea.org/weo2017/. [9] 2018 Snapshot of Global Photovoltaic Markets. http://www.iea-pvps.org/index.php?id¼266. [10] International Energy Agency, World Energy Outlook 2016, Organization for Economic Cooperation and Development (OECD)/IEA, Paris, 2016, ISBN 978-92-64-26495-3. OCLC: 963552636. [11] D.M. Chapin, C.S. Fuller, G.L. Pearson, A new silicon pn junction photocell for converting solar radiation into electrical power, Journal of Applied Physics 25 (5) (1954) 676. https://doi. org/10.1063/1.1721711. http://aip.scitation.org/ doi/10.1063/1.1721711. [12] H.S. Rauschenbach, Solar Cell Array Design Handbook, 978-94-011-7917-1 978-94-0117915-7, Springer, Dordrecht, 1980. https://doi. org/10.1007/978-94-011-7915-7_4, https://link. springer.com/chapter/10.1007/978-94-0117915-7_4. [13] L. Fraas, L. Partain, Solar cells: a brief history and introduction, in: Solar Cells and Their Applications, Wiley-Blackwell, 2010, ISBN 978-0-47063688-6, pp. 1e15. https://doi.org/10.1002/ 9780470636886.ch1. https://onlinelibrary.wiley. com/doi/abs/10.1002/9780470636886.ch1. [14] R.L. Easton, M.J. Votaw, Vanguard I IGY satellite (1958 beta), Review of Scientific Instruments 30 (1959) 70e75. https://doi.org/10. 1063/1.1716492. [15] H.E. Newell, Scientific Findings From Explorer VI, Technical Report NASA-SP-54: 1965001 2364, NASA, 1965, https://ntrs.nasa.gov/ search.jsp?R¼19650012364. [16] D.M. Harland, R. Lorenz, Space Systems Failures: Disasters and Rescues of Satellites, Rocket and Space Probes, Springer Science & Business Media, 2007.

PV S YSTEMS

17

[17] A.W. Czanderna, F.J. Pern, Encapsulation of PV modules using ethylene vinyl acetate copolymer as a pottant: a critical review, Solar Energy Materials and Solar Cells 43 (2) (1996) 101e181. https://doi.org/10.1016/0927-0248 (95)00150-6. http://www.sciencedirect.com/ science/article/pii/0927024895001506. [18] C. Peike, I. Ha¨drich, K.-A. Weiß, Overview of PV module encapsulation materials, Photovoltaics International 19 (2013) 85e92. PV Tech, https:// www.pv-tech.org/technical-papers/overview-ofpv-module-encapsulation-materials. [19] E.F.C. Cuddihy, Applications of Ethylene Vinyl Acetate as an Encapsulation Material for Terrestrial Photovoltaic Modules, 1983 tech. rep. https:// ntrs.nasa.gov/search.jsp?R¼19830021538. [20] E. Christensen, Electricity from photovoltaic solar cells, in: Flat-Plate Solar Array Project of the US Department of Energy’s National Photovoltaics Program: 10 Years of Progress, 1985 tech. rep. https://ntrs.nasa.gov/search.jsp? R¼19880002806. [21] M.I. Smokler, D.H. Otth, R.G. Ross, The block program approach to photovoltaic module development, in: Proceedings of the 18th IEEE PV Specialists Conference, 1985, pp. 1150e1158. [22] E. Cuddihy, C. Coulbert, A. Gupta, R. Liang, Electricity from Photovoltaic Solar Cells: FlatPlate Solar Array Project Final Report. Volume VII: Module Encapsulation, 1986. http:// resolver.caltech.edu/JPLpub86-31-volumeVII. [23] A. Rosenthal, M. Thomas, S. Durand, A 10-year review of performance of photovoltaic systems, in: Conference Record of the Twenty Third IEEE Photovoltaic Specialists, 1993, pp. 1289e1291. [24] D.C. Jordan, S.R. Kurtz, Photovoltaic degradation ratesan analytical review, Progress in Photovoltaics: Research and Applications 21 (1) (2013) 12e29. https://doi.org/10.1002/pip.1182. http://onlinelibrary.wiley.com/doi/10.1002/pip. 1182/abstract. [25] H.J. Wenger, J. Schaefer, A. Rosenthal, B. Hammond, L. Schlueter, Decline of the Carrisa Plains PV power plant: the impact of concentrating sunlight on flat plates, in: The Conference Record of the Twenty-Second IEEE Photovoltaic Specialists Conference e 1991, vol. 1, 1991, pp. 586e592. https://doi.org/10. 1109/PVSC.1991.169280.

18

D URABILITY AND R ELIABILITY

OF

P OLYMERS

[26] J.C. Arnett, L.A. Schaffer, J.P. Rumberg, R.E.L. Tolbert, Design, installation and performance of the ARCO Solar one-megawatt power plant, in: 5th Photovoltaic Solar Energy Conference, Athens Greece: Dordrecht, D, Reidel Publishing Co, 1984, pp. 314e320. http:// adsabs.harvard.edu/abs/1984pvse.conf..314A. [27] C. Gay, E. Berman, Performance of large photovoltaic systems, Chemtech 20 (3) (1990) 182e186. http://apps.isiknowledge.com/full_ record.do?product¼WOS&search_mode¼General Search&qid¼19&SID¼3CH51D78eh5llM1h GdB&page¼1&doc¼1. [28] H.J. Wenger, C. Jennings, J.J. Iannucci, Carrisa Plains PV power plant performance, in: IEEE Conference on Photovoltaic Specialists, vol. 2, 1990, pp. 844e849. https://doi.org/10.1109/ PVSC.1990.111740. [29] A.L. Rosenthal, C.G. Lane, Field-test results for the 6 Mw Carrizo solar photovoltaic powerplant, Solar Cells 30 (1e4) (1991) 563e571. http://apps.isiknowledge.com/full_record.do? product¼WOS&search_mode¼GeneralSearch &qid¼4&SID¼3CH51D78eh5llM1hGdB&page ¼1&doc¼8. [30] J. Schaefer, L. Schlueter, A. Rosenthal, H. Wenger, Electrical degradation of the Carrisa Plains power plant, in: Tenth E.C. Photovoltaic Solar Energy Conference, Springer, Dordrecht, 1991, pp. 1248e1253. https://doi.org/10.1007/ 978-94-011-3622-8_315. ISBN: 978-94-0105607-6, 978-94-011-3622-8, https://link.springer. com/chapter/10.1007/978-94-011-3622-8_315. [31] D.D. Sumner, C.M. Whitaker, L.E. Schlueter, Carrisa Plains photovoltaic power plant 1984e7 performance, in: Conference Record of the Twentieth IEEE Photovoltaic Specialists Conference, vol. 2, 1988, pp. 1289e1292. https:// doi.org/10.1109/PVSC.1988.105912. [32] T. Hoff, G. Shushnar, Two years of performance data for the world’s largest photovoltaic power plant, IEEE Transactions on Energy Conversion EC-2 (2) (1987) 232e235. https://doi.org/10. 1109/TEC.1987.4765835. [33] J.H. Wohlgemuth, R.C. Petersen, Reliability of EVA modules, in: Conference Record of the Twenty Third IEEE Photovoltaic Specialists Conference e 1993 (Cat. No.93CH3283-9), 1993, pp. 1090e1094. https://doi.org/10.1109/ PVSC.1993.346972.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

[34] J.C. Schaefer, Review of photovoltaic power plant performance and economics, IEEE Transactions on Energy Conversion 5 (2) (1990) 232e238. https://doi.org/10.1109/60.107215. [35] S. Hester, PVUSA-advancing the state-of-the-art in photovoltaics for utility scale applications, in: Conference Record of the Twentieth IEEE Photovoltaic Specialists Conference, vol. 2, 1988, pp. 1068e1074. https://doi.org/10.1109/ PVSC.1988.105869. [36] S. Hester, T. Townsend, W. Clements, W. Stolte, PVUSA: lessons learned from startup and early operation, in: IEEE Conference on Photovoltaic Specialists, Kissimmee, FL, USA, IEEE, 1990, pp. 937e943. https://doi.org/10.1109/PVSC. 1990.111757. http://ieeexplore.ieee.org/document/ 111757/. [37] R. Steele, PVUSA Progress Report, 1989e1990, tech. rep., Pacific Gas and Electric Co., San Ramon, CA (United States), 1990, 00000, http:// www.osti.gov/scitech/biblio/10163193. [38] T.R. Candelario, S.L. Hester, T.U. Townsend, D.J. Shipman, PVUSA-performance, experience, and cost (PV power systems), in: The Conference Record of the Twenty-Second IEEE Photovoltaic Specialists Conference e 1991, vol. 1, 1991, pp. 493e500. https://doi.org/10. 1109/PVSC.1991.169264. [39] W. Ellyn, C. Jennings, PVUSA Progress Report, 1991, tech. rep., Pacific Gas and Electric Co, San Ramon, CA (United States), 1991, 00000, http:// www.osti.gov/scitech/biblio/10163200. [40] A. Luque, Will we exceed 50% efficiency in photovoltaics? Journal of Applied Physics 110 (3) (2011) 031301. https://doi.org/10.1063/1. 3600702. http://aip.scitation.org/doi/abs/10. 1063/1.3600702. [41] M.J. O’Neill, A.J. McDanal, R.R. Walters, J.L. Perry, Recent developments in linear Fresnel lens concentrator technology, including the 300 kW 3m/Austin system, the 20 kW PVUSA system, and the concentrator initiative, in: The Conference Record of the Twenty-Second IEEE Photovoltaic Specialists Conference e 1991, vol. 1, 1991, pp. 523e528. https://doi.org/10. 1109/PVSC.1991.169269. [42] R.N. Dows, E.J. Gough, PVUSA Model Technical Specification for a Turnkey Photovoltaic Power System, tech. rep., Pacific Gas and Electric Co., San Ramon, CA (United States), 1995.

1: OVERVIEW: P OWER D EGRADATION

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

AND

FAILURE

OF

Dept. of Research and Development, http:// www.osti.gov/scitech/biblio/172103. R.N. Dows, E.J. Gough, PVUSA Procurement, Acceptance, and Rating Practices for Photovoltaic Power Plants, Tech. Rep. DOE/AL/8299321, Pacific Gas and Electric Co., San Ramon, CA (United States), Dept. of Research and Development; Bechtel Corp., San Francisco, CA (United States), 1995. https://doi.org/10.2172/ 119944, https://www.osti.gov/biblio/119944. C. Jennings, B. Farmer, T. Townsend, P. Hutchinson, T. Reyes, C. Whitaker, et al., PVUSA-the first decade of experience, in: Conference Record of the Twenty Fifth IEEE Photovoltaic Specialists Conference e 1996, 1996, pp. 1513e1516. https://doi.org/10.1109/ PVSC.1996.564424. C.M. Whitaker, T.U. Townsend, J.D. Newmiller, D.L. King, W.E. Boyson, J.A. Kratochvil, et al., Application and validation of a new PV performance characterization method, in: Conference Record of the Twenty Sixth IEEE Photovoltaic Specialists Conference e 1997, 1997, pp. 1253e1256. https://doi.org/10.1109/PVSC. 1997.654315. D.R. Myers, Evaluation of the Performance of the PVUSA Rating Methodology Applied to Dual Junction PV Technology: Preprint (Revised), Tech. Rep. NREL/CP-550-45376, National Renewable Energy Lab. (NREL), Golden, CO (United States), 2009, https://www. osti.gov/biblio/951223. T. Townsend, P. Hutchinson, Soiling analyses at PVUSA, in: Proceedings of the Solar Conference, American Solar Energy Society; American Institute Of Architects, 2000, pp. 417e422. International Technology Roadmap for Photovoltaic (ITRPV) e Reports/Downloads. http:// www.itrpv.net/Reports/Downloads/. International Energy Agency, World Energy Outlook 2018, Organization for Economic Cooperation and Development (OECD)/IEA, Paris, France, 2018, ISBN 978-92-64-30677-6. R. Scott, The History of the International Energy Agency e the First 20 Years, vol. 1, 1994. https:// www.oecd-ilibrary.org/content/publication/9789 264020931-en. R. Scott, The History of the International Energy Agency: Major Policies and Actions, vol. 2, 1994. https://www.oecd-ilibrary.org/energy/

PV S YSTEMS

[52]

[53]

[54]

[55]

[56]

[57]

[58]

[59]

[60]

[61]

19

the-history-of-the-international-energy-agency_ 9789264021006-en. R. Scott, The History of the International Energy Agency: Principle Documents, vol. 3, 1995. https://www.oecd-ilibrary.org/energy/the-historyof-the-international-energy-agency_978926402 1006-en. C.S. Bamberger, The History of the International Energy Agency: Supplement to Volumes I, II, III, vol. 4, 2004. https://www.oecd-ilibrary.org/ energy/the-history-of-the-international-energyagency_9789264014947-en. M.J. Brunisholz, Stefan Nowak, 2018. ieapvps.org-Who we are, http://www.iea-pvps.org/ index.php?id¼4#c38. U. Jahn, B. Farnung, iea-pvps.org Task 13 Contacts, 2018. http://www.iea-pvps.org/index.php? id¼61. T. Nordmann, L. Clavadetscher, W. GJHM van Sark, M. Green, Analysis of Long-Term Performance of PV Systems, Tech. Rep. IEA-PVPS T13-05, 2014, 2014, http://iea-pvps.org/index. php?id¼57&eID¼dam_frontend_push&doc ID¼2306. M. Ko¨ntges, S. Kurtz, C. Packard, U. Jahn, K.A. Berger, K. Kato, T. Friesen, et al., Review of Failures of Photovoltaic Modules, 2014. IEA PVPS Task 13. tech. rep. http://www.iea-pvps. org/index.php?id¼275. M. Ko¨ntges, G. Oreski, U. Jahn, M. Herz, P. Hacke, K.-A. Weiss, et al., IEA-PVPS Task 13, SubTask 3.4: Assessment of Photovoltaic Module Failures in the Field, T13-09-2017, IEAPVPS Task 13, Tech. Rep, 2017, http://www. iea-pvps.org/index.php?id¼435. U. Jahn, M. Herz, M. Ko¨ntges, D. Parlevliet, M. Paggi, I. Tsanakas, et al., Review on IR and EL Imaging for PV Field Applications, Tech. Rep. IEA-PVPS T13-10:2018, IEA-PVPS Task 13, 2018, http://www.iea-pvps.org/index.php?id¼480. M. Richter, C. Tjengdrawira, J.V. SiCon, M. Green, L. Frearson, B. Herteleer, et al., Technical Assumptions Used in PV Financial Models e Review of Current Practices and Recommendations, 2017. Tech. Rep. T13-082017, IEA-PVPS Task 13, http://www.ieapvps.org/index.php?id¼426. N. Marigo, The Chinese silicon photovoltaic industry and market: a critical review of trends and outlook, Progress in Photovoltaics: Research

20

[62]

[63]

[64]

[65]

[66]

[67]

D URABILITY AND R ELIABILITY

OF

P OLYMERS

and Applications 15 (2) (2007) 143e162. https:// doi.org/10.1002/pip.716. https://onlinelibrary. wiley.com/doi/abs/10.1002/pip.716. J. Chang, D.Y.C. Leung, C.Z. Wu, Z.H. Yuan, A review on the energy production, consumption, and prospect of renewable energy in China, Renewable and Sustainable Energy Reviews 7 (5) (2003) 453e468. https://doi.org/10.1016/S13640321(03)00065-0. http://www.sciencedirect.com/ science/article/pii/S1364032103000650. A. de la Tour, M. Glachant, Y. Me´nie`re, Innovation and international technology transfer: the case of the Chinese photovoltaic industry, Energy Policy 39 (2) (2011) 761e770. https://doi. org/10.1016/j.enpol.2010.10.050. http://www. sciencedirect.com/science/article/pii/S0301421 510008013. H. Sun, Q. Zhi, Y. Wang, Q. Yao, J. Su, Chinas solar photovoltaic industry development: the status quo, problems and approaches, Applied Energy 118 (Supplement C) (2014) 221e230. https://doi.org/10.1016/j.apenergy.2013.12.032. http://www.sciencedirect.com/science/article/pii/ S0306261913010313. S. Zhang, P. Andrews-Speed, X. Zhao, Y. He, Interactions between renewable energy policy and renewable energy industrial policy: a critical analysis of China’s policy approach to renewable energies, Energy Policy 62 (2013) 342e353. https://doi.org/10.1016/j.enpol.2013. 07.063. http://www.sciencedirect.com/science/ article/pii/S0301421513007064. S. Zhang, Y. He, Analysis on the development and policy of solar PV power in China, Renewable and Sustainable Energy Reviews 21 (2013) 393e401. https://doi.org/10.1016/j.rser.2013. 01.002. http://www.sciencedirect.com/science/ article/pii/S1364032113000269. H. Zou, H. Du, J. Ren, B.K. Sovacool, Y. Zhang, G. Mao, Market dynamics, innovation, and transition in China’s solar photovoltaic (PV) industry: a critical review, Renewable and Sustainable Energy Reviews 69 (2017) 197e206. https://doi.org/10.1016/j.rser.2016. 11.053. http://www.sciencedirect.com/science/ article/pii/S1364032116308176.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

[68] V. Devabhaktuni, M. Alam, S.S.S.R. Depuru, R.C. Green, D. Nims, C. Near, Solar energy: trends and enabling technologies, Renewable and Sustainable Energy Reviews 19 (2013) 555e564. https://www.sciencedirect.com/ science/article/pii/S1364032112006363. [69] E.D. Dunlop, D. Halton, H.A. Ossenbrink, 20 years of life and more: where is the end of life of a PV module?, in: Conference Record of the Thirty-First IEEE Photovoltaic Specialists Conference, 2005, 2005, pp. 1593e1596. https:// doi.org/10.1109/PVSC.2005.1488449. [70] J.H. Wohlgemuth, D.W. Cunningham, P. Monus, J. Miller, A. Nguyen, Long term reliability of photovoltaic modules, in: 2006 IEEE 4th World Conference on Photovoltaic Energy Conference, vol. 2, 2006, pp. 2050e2053. https://doi.org/10. 1109/WCPEC.2006.279905, vol. 2. [71] CSI, Canadian Solar: Warranty Overview for Canadian Solar Modules, 2018. https://www. canadiansolar.com/warranty.html. [72] J.L. Prince, J.W. Lathrop, F.W. Morgan, E.L. Royal, G.W. Witter, Accelerated stress testing of terrestrial solar cells, in: 17th International Reliability Physics Symposium, 1979, pp. 77e86. https://doi.org/10.1109/IRPS.1979.362874. [73] J.L. Prince, J.W. Lathrop, Investigation of Reliability Attributes and Accelerated Stress Factors of Terrestrial Solar Cells, Second quarterly report, March 9, 1978eJune 9, 1978. Tech. Rep. DOE/JPL/954929-2, Clemson Univ., SC (USA), 1978. https://doi.org/10.2172/6642723, https:// www.osti.gov/biblio/6642723. [74] J.W. Lathrop, J.L. Prince, Investigation of Reliability Attributes and Accelerated Stress Factors of Terrestrial Solar Cells, Second annual report. Tech. Rep. DOE/JPL/954929-80/7, Clemson Univ., SC (USA), 1980. https://doi.org/ 10.2172/5595307, https://www.osti.gov/biblio/ 5595307. [75] L.S. Bruckman, N.R. Wheeler, J. Ma, E. Wang, C.K. Wang, I. Chou, et al., Statistical and domain analytics applied to PV module lifetime and degradation science, IEEE Access 1 (2013) 384e403. https://doi.org/10.1109/ACCESS. 2013.2267611.

1: OVERVIEW: P OWER D EGRADATION

AND

FAILURE

OF

[76] R.H. French, P. Rudolf, T.J. Peshek, L.S. Bruckman, Y. Xu, N.R. Wheeler, et al., Degradation science: mesoscopic evolution and temporal analytics of photovoltaic energy materials, Current Opinion in Solid State and Materials Science 19 (4) (2015) 212e226. https://doi.org/10.1016/j.cossms.2014.12.008. http://www.sciencedirect.com/science/article/ pii/S1359028614000989. [77] T. Surek, Progress in U.S. photovoltaics: looking back 30 years and looking ahead 20, in: Proceedings of 3rd World Conference onPhotovoltaic Energy Conversion, 2003, vol. 3, 2003, pp. 2507e2512. [78] N.G. Dhere, Reliability of PV modules and balance-of-system components, in: Conference Record of the Thirty-First IEEE Photovoltaic Specialists Conference, 2005, 2005, pp. 1570e1576. https://doi.org/10.1109/PVSC. 2005.1488445. [79] tech. rep., EIA e Annual Energy Outlook 2018, 2018. https://www.eia.gov/outlooks/aeo/. [80] The General Assembly, Transforming Our World: The 2030 Agenda for Sustainable Development, 2015. http://www.un.org/ga/search/ view_doc.asp?symbol¼A/RES/70/1&Lang¼E.

PV S YSTEMS

21

[81] International Energy Agency, World Energy Model, Tech. Rep. WEM2018, International Energy Agency, 2018, https://www.iea.org/weo/ weomodel/. [82] M. Metayer, C. Breyer, H.-J. Fell, The Projections for the Future and Quality in the Past of the World Energy Outlook for Solar PV and Other Renewable Energy Technologies, tech. rep., Energy Watch Group, Berlin, Germany, 2015. http:// energywatchgroup.org/wp-content/uploads/2018/ 01/EWG_WEO-Study_2015.pdfs_.pdf. [83] A. Ja¨ger-Waldau, Snapshot of photovoltaicsmarch 2017, Sustainability 9 (5) (2017) 783. https://doi.org/10.3390/su9050783. https://www. mdpi.com/2071-1050/9/5/783. [84] A. Ja¨ger-Waldau, Snapshot of photovoltaics e february 2018, EPJ Photovoltaics 9 (2018) 6. https://doi.org/10.1051/epjpv/2018004. https:// www.epj-pv.org/10.1051/epjpv/2018004. [85] F. Creutzig, P. Agoston, J.C. Goldschmidt, G. Luderer, G. Nemet, R.C. Pietzcker, The underestimated potential of solar energy to mitigate climate change, Nature Energy 2 (9) (2017) 17140. https://doi.org/10.1038/nenergy. 2017.140. https://www.nature.com/articles/ nenergy2017140.

2 Overview: Durability and Reliability of Common PV Module and Polymers/Materials Hsinjin Edwin Yang Pioneer Scientific Solutions, LLC., Long Grove, IL, United States

2.1 Background Concerns about global warming have led to increasing efforts to reduce carbon dioxide emissions. Motivated by change in public policy, utilities, corporations, and individual homeowners are increasingly adding renewable sources of energy that do not use fossil fuels for generating electricity. Fig. 2.1 shows the recent increase in the US shipments of photovoltaic systems [1]. The global cumulative solar photovoltaic (PV) modules/panels installed capacity increased significantly from about 40 GW in 2010 to 320 GW in 2016, at a Compound Annual Growth Rate (CAGR) of 40% between 2010 and 2016. During the 2012e20 forecast period, solar PV installations will experience a steady growth globally, and the cumulative installed capacity is expected to reach 329.8 GW in 2020, growing at a CAGR of 18.5% from 2011 to 2020. For photovoltaic systems to be competitive with traditional (e.g., fossil, hydro) and other renewable energy sources (e.g., wind), they need to perform with reliability over a long duration. Thus one longterm goal of the U.S. Department of Energy’s Solar

Energy Technology Office is to develop costeffective PV modules with 30-year useful lifetimes. To meet this goal there is a strong impetus to understand the performance of packaging materials that would support the desired long-term photovoltaic electric power generation in a system. These materials include glazing, polymers, and metals used in making photovoltaic systems. Of particular interest are polymeric materials as these may degrade from ultraviolet radiation, diurnal/annual temperature and humidity swings, air pollution, water ingress, and salty conditions leading to performance (e.g., drop in system efficiency) and safety issues (electrical and fire hazards). For example, polymeric materials are used to encapsulate the photovoltaic cells as a barrier to the environment (e.g., rain, salt, and dust). The materials on the collection side of a module need to be transparent in the radiation wavelength range that the photovoltaic system uses to generate electricity. They also need to have a certain dielectric strength and arc resistance to prevent electrical and fire hazards respectively. However, changes in the chemical structure due to ultraviolet exposure, temperature, humidity, and/or the combination of these stressors

160 140

Peak (105 KW)

120 100 80 60 40

20 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Year Figure 2.1 US photovoltaic shipments in peak power (kW) generated (2006e16). Durability and Reliability of Polymers and Other Materials in Photovoltaic Modules. https://doi.org/10.1016/B978-0-12-811545-9.00002-1 Copyright © 2019 Hsinjin Edwin Yang. Published by Elsevier Inc. All rights reserved.

23

24

D URABILITY AND R ELIABILITY

OF

P OLYMERS

may reduce their light transmission properties and also their dielectric strength and arc ignition resistance. These changes subsequently may have an adverse effect on the photovoltaic system reliability. Similarly, degradation of other polymeric materials used in PV systems (e.g., substrate, electrical enclosures, and wiring) may also lead to performance and safety issues over the life of the system. Test standards are available to assess the performance and safety attributes of photovoltaic modules and panels. Standards for newer concentrated photovoltaic systems (CPV) are also under development by ASTM and IEC technical committees. While these standards provide considerable attention to panel and system performance, they do not provide adequate focus on safety issues. For example, a number of these systems have malfunctioned, resulting in fires. Some of these incidents have resulted from cell failures and electric arcing. In the United States, there is also a concern that solar panels may be an additional fuel source that may compromise the fire performance of roof systems. The current safety standards (IEC 61730 [2] and UL 1703 [3]) for polymeric materials used in PV modules are summarized in Appendix I. There are some issues in the current standards, such as the aging of polymeric materials which is addressed by thermal oxidation via the relative thermal index (RTI). In addition, these standards use existing methodology on thermal aging of polymeric materials where they are exposed to ultraviolet exposure, temperature, or humidity individually. This has now been recognized, and the proposed revision to IEC 61730, and UL 1703 will include provisions for damp heat exposure that combine temperature and humidity, but still do not include ultraviolet radiation. Thus, these referenced methods do not represent the actual field conditions (UV, temperature, humidity, and the combination of these stressors) that panels are exposed to. Further, there is a lack of understanding regarding how much exposure time and strength applied on the polymeric materials will cause failure and how various environmental factor(s) correlate with or contribute to electrical and safety hazards of PV modules. Data driven approaches that encompass a large-set of data from fielded PV modules will be useful to understand the degradation of these materials in PV modules under multistress conditions [4,5]. Recent results from visual inspections and studies [6e8] on used PV modules indicated that most

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

defects and/or failures arising from material-related issues are: 1. discoloration of encapsulant and substrate, 2. oxidation or browning of the cell electrical grid, 3. encapsulant and backsheet delamination, 4. substrate delamination, 5. junction box detachment, and 6. glass cracking.

2.2 Fundamentals for Durability and Reliability of PV Materials 2.2.1 Basic Structure and Materials for a PV Module The basic structure of a major PV module with crystal silicon (CeSi) is shown in Fig. 2.2, and it is composed of: (A) a front sheet or superstrate of mainly glass; (B) an encapsulant, mainly of ethylene vinyl acetate (EVA), but there are some recently developed and improved polymers, such as polyolefin elastomer (POE), Polyvinylbutyral (PVB) hydrocarbon ionomers, and organo-silicon; (C) a cell mainly of crystal silicon (CeSi) or some amorphous silicon; (D) a substrate mainly a composite film of PVF/PET/EVA or PVF/PET/PVF, where PVF: polyvinylidene fluoride and PET: polyethylene terephthalate, and (E) a junction box and structural parts mainly of PET, acrylonitrile butadiene styrene (ABS), or polycarbonate (PC). The types of modules and materials are listed in Table 2.1 and the primary functions and requirements of the materials used in PV superstrate, encapsulant, and substrate components are listed in Table 2.2. It is important to mention here that PERC (passivated emitter rear cell) was first developed in Australia in the 1980s by scientist Martin Green and his team at the University of New South Wales, and the PERC technology adds an extra layer to the rearside of a solar cell. Manufacturers spent many years focusing on the front side of a solar cell, and less attention was paid to taking advantage of production opportunities from the backside (Fig. 2.3). Incorporating a PERC into a solar cell boosts generation. In order to create a PERC cell, additional two steps are employed to the standard back surface field (BSF) during the manufacturing process. First, a rear

2: OVERVIEW: D URABILITY AND R ELIABILITY

OF

C OMMON PV M ODULE

AND

P OLYMERS /M ATERIALS

25

A B C B F E

H 3C O C H O C C

H H C C H H

n

H H

m

H

F

C

C

H

H

*

O

O

O

C

C

O

CH2

CH2

* n

n

Figure 2.2 The basic structure of a PV mini-module with the basic polymers for the encapsulant (EVA, left side) and substrate (PVFdcenter and PETdright). Table 2.1 Types of PV Modules and Materials List PV Modules/Materials List Module Type

Front Sheet (Superstrate)

Encapsulant

Backsheet (Substrate)

Cell (Module) Efficiency

Market Trend

Glass/CeSi/ Polymer

Glass

EVA, PVB, POE, ionomers, organo-silicon

Fluoropolymer Multilayers

17%e20% Max. 23.0%

Major

Glass/TF/Glass

Glass

EVA, PVB, POE, ionomers, organo-silicon

Glass

10%e12% Max. 14.4%

Minor

Fluoropolymers

EVA

Fluoropolymer Multilayers

Next

Glass

EVA, PVB, organo-silicon

Fluoropolymer Multilayers

Next

Fluoropolymers

EVA, PVB, organo-silicon

Fluoropolymer Multilayers

Next

Polymer/Ce Si/Polymer Glass/TF/ Polymer Polymer/TF/ Polymer

CdTe, cadmium telluride; CIGS, coppereindiumegalliumeselenide; TF, thin film.

surface passivation film is applied. Second, lasers or chemicals are used to open the rear passivation stack and create tiny pockets in the film to absorb more light. In employing just two additional steps, there are three advantages: (1) the electron recombination is significantly reduced; (2) more light is absorbed; and (3) a higher internal reflectivity is experienced. Not all sunlight is absorbed through non-PERC solar cells since some light passes straight through.

Applying with a passivation layer on the rear side of a PERC cell, unabsorbed light is reflected by the additional layer back to the solar cell for a second absorption attempt. This process leads to a more efficient solar cell. This is great news for those across the spectrum of the industry. The PERC technology has been increasingly applied in recent years because it does not require complete operation overhauls by current cell manufacturers.

26

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Table 2.2 PV Module Material Functional Requirements PV Components Superstrate (front sheet)

Functions/Requirements
Mechanical rigidity and impact resistance
Good adhesion to encapsulant
Optical transparency

Current Materials Glass: tempered/low-iron (Si); non-tempered/low-iron (thinfilm)
Polymeric film: fluoropolymers: ETFE


Electrical isolation of the solar cell circuits
Water/moisture barrier
Weathering durability Encapsulant


Adequate mechanical compliance to accommodate stresses
High optical transmittance

Ethylene-vinyl-acetate (EVA) Polyvinyl butyral (PVB) Silicon rubber


Good adhesion to different module materials
Good dielectric properties (electrical insulation) Substrate (back sheet)


Mechanical strength
Good adhesion to encapsulant
Water/moisture barrier

PVF/PET/PVF PVF/PET/EVA Transparent polymer film: ETFE, PCTFE


Electrical insulation
Weathering durability
Optical transparency

Figure 2.3 The detailed diagram of PERC cell technology compared with a conventional monocrystalline cell. By Kelly Pickerel, Solar Power World (2016).

2.2.2 Functional Requirements of PV Module Materials To survive in harsh operating environments, PV modules rely on packaging materials including protective superstrate, substrate, sealants, and

encapsulants to provide requisite durability. Several key properties associated with PV module durability are critical for commercial success. These include low-interface conductivity, adequate adhesion of encapsulants to substrate, superstrate, and PV cells, low moisture permeation through all packaging

2: OVERVIEW: D URABILITY AND R ELIABILITY

OF

C OMMON PV M ODULE

materials, and excellent mechanical properties such as tensile elongation, creep resistance, and impact strength at all operating conditions. Based on the published results [7e9] and UL’s experiences, it has been demonstrated that accelerated damp-heat exposure (85  C, 85% RH for 1000 h) is a severe condition and would be considered as a convenient screening tool for selecting the PV materials and continuing the study. It is noted that the damp heat may be an over-accelerating degradation test, and is not actually a good test especially for materials with Tg much lower than 85  C. Therefore, the damp heat should not be used as a qualification test for PV modules. The primary functions and requirements of the materials used in PV superstrate (front sheet), encapsulant, and substrate (backsheet) components are listed in Table 2.2 and described in the following.

2.2.2.1 Superstrate (Front Sheet) The front surface of a PV module must have a high transmission in the wavelengths which can be used by the solar cells in the PV module. For silicon solar cells, the top surface must have high transmission of light in the wavelength range of 350e1200 nm. In addition, the reflection from the front surface should be low. While theoretically this reflection could be reduced by applying an antireflection coating to the top surface, in practice these coatings are not robust enough to withstand the conditions in which most PV systems are used. An alternative technique to reduce reflection is to “roughen” or texture the surface. However, in this case the dust and dirt is more likely to attach itself to the top surface, and less likely to be dislodged by wind or rain. These modules are not therefore “self-cleaning,” and the advantages of reduced reflection are quickly outweighed by losses incurred due to increased top surface soiling. In addition to its reflection and transmission properties, the top surface material should be impervious to water, should have good impact resistance, should be stable under prolonged UV exposure, and should have a low thermal resistivity. Water or water vapor ingress into a PV module will corrode the metal contacts and interconnects, and consequently will dramatically reduce the lifetime of the PV module. In most modules, the front surface is used to provide the mechanical strength and rigidity, therefore either the top surface or the rear surface

AND

P OLYMERS /M ATERIALS

27

must be mechanically rigid to support the solar cells and the wiring. There are several choices for a top surface material including acrylic, polymers, transparent fluoropolymers, and glass. Tempered, low iron-content glass is most commonly used as it is of low cost, strong, stable, highly transparent, and impervious to water and gases and has good self-cleaning properties. However, the glass has disadvantages of easy breakage and heavy weight. Corning Incorporated has recently produced photovoltaic (PV) Willow and/ or Gorilla glass for superstrates and/or substrates with stronger impact, higher efficiency, and lighter weight thin-film photovoltaic modules.

2.2.2.2 Encapsulants An encapsulant is used to provide adhesion between the solar cells, the top surface, and the rear surface of the PV module. The encapsulant should be stable at elevated temperatures and high UV exposure. It should also be optically transparent and should have a low thermal resistance. EVA (ethylene vinyl acetate) is the most commonly used encapsulant material. EVA comes in thin sheets which are inserted between the solar cells and the top surface and the rear surface. This composite is then heated to 150 C to polymerize the EVA and bond the module together. However, some material/chemical companies have recently developed new encapsulating materials for faster, more efficient module production and longer service life for both crystalline silicon and thin-film PV modules. For example, Arkema Co. and Wacker Chemie introduced a highly transparent nanostructured thermoplastic polymerdApolhya Solar and a thermoplastic-organo-silicon elastomerdTectosil for use without a curing process, respectively. DuPont recently developed PV5200 series encapsulant sheets based on polyvinyl butyral (PVB) polymer technology and PV5400 ionomers based encapsulants, and Dow ENGAGE PV Polyolefin Elastomers (POEs) especially for the needs of growing thin-film module technology and market with improving power efficiency. They all claimed that the above encapsulant materials show excellent glass adhesion, proven safety glass performance with high visible light transmission, increased protection from moisture, and improved power retention.

28

D URABILITY AND R ELIABILITY

OF

P OLYMERS

2.2.2.3 Substrate (Backsheet) The key characteristics of the rear surface or substrate of the PV module include both low thermal resistance and preventing the ingress of water or water vapor. In most modules, a thin polymer sheet, typically PVF or Tedlar®, is used as the rear surface. In order to gain good adhesion and moisture barrier with reasonable cost, fluoropolymers and their multilayered film with other polymers (e.g., PET and EVA) or metals (e.g., Al foil) are the most popular materials for backsheet. The most popular backsheets used in the PV module are PVF/PET/PVF and PVF/ PET/EVA. Please note that there is an additional important criterion, like high light transmission for both front sheet of PV modules and backsheet in the case of bifacial ones, and interlayer adhesion for multilayer sheets. Very recently, DuPont Photovoltaic Solutions has launched the first transparent PV module backsheet material, specifically for bifacial solar modules at SNEC International Photovoltaic Power Generation and Smart Energy Exhibition, being held in Shanghai, China during the week of May 27, 2018. DuPont noted that such clear “Tedlar” PVF film based backsheets allow higher reliability, lower operating temperature, compared to glass/glass bifacial modules, as well as being 30% lighter. The most important effects in PV modules or arrays [5e8] are: 1. losses due to the interconnection of mismatched solar cells, 2. environmental exposure of the modules, 3. failure modes of PV modules, 4. lifetime and durability of PV modules, and 5. safety of PV modules for long-term exposure.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

The durability and reliability of the materials are very critical to the above effects. The relevant main theories for describing the environmental and aging processes are briefly outlined in the following section.

2.2.3 Reliability Versus Durability We are generally looking for or purchasing a product with warranted reliability. Reliability can have many definitions, but in a broad sense it is the measure of a product (or service), and it is primarily concerned with discrete, absolute failure including time to failure, mean time between failures, and percentage of population failed, etc. of the overall system as functions of temperature (T), humidity (H), radiation (UV), and other internal and external stresses such as pressure (P) and impact. Durability usually involves understanding the route to failure (mechanisms) and the property rate of change (kinetics). These individual changes may not lead to loss of reliability but may result in declining performance and shortened service lifetimes. Durability contributes to reliability which is partially a function of durability as shown in Fig. 2.4. As we know, the photovoltaic (PV) module and system are comprised of complex integration of front and back sheets, encapsulant, photocells, cables, connectors, sealant, adhesive and junction box, etc. and they are used for long term service life preferably more than 25 years. The reliability of a PV product and system will rely on the durability of the individual materials and/or component durability. Therefore, the measurements for investigating reliability and understanding durability are very essential to assure the safety and quality of the PV module and system. Reliability measurements typically include failure rates, cumulative failures, component

Figure 2.4 The comparison between reliability (left) and durability (right).

2: OVERVIEW: D URABILITY AND R ELIABILITY

OF

C OMMON PV M ODULE

lifetimes, and estimates of product lifetimes. Durability measurements may include changes to chemical, physical, or appearance properties, loss of performance, rate of property change with time or stress, and time to unacceptable performance. The objectives of durability and reliability studies and understandings are to assure the safety and quality of a PV module and its system.

2.2.4 Theoretical Background for the Durability and Reliability Study 2.2.4.1 Solar Irradiance [10,11] The amount of solar irradiance received by a panel is measured as follows: Irradiance ¼ GHI 

cos½tilt cos½zenith

2.2.4.2 Thermal Process [10,11] The module temperature (Tm) is estimated from the irradiance and the wind speed (WS) by Eq. (2.2): Tm ¼ Tamb þ Irradiance (2.2)

where Tamb is the ambient temperature. Thermal expansion and thermal stresses can be described by the following equation [11]: d ¼ ðaG C  aC DÞDT

P OLYMERS /M ATERIALS

29

crosslinking and the chain scission reaction yields of polymers being irradiated by ionizing radiation. Such equation can be modified [13] for simultaneous UV crosslinked and degraded polymers by simple replacement of dose with the aging time, sþ

  pffiffi s ¼ p0 =q0 þ 2 q0 u2;0 t

(2.4)

where t is the aging time, s is the sol fraction, p0 is the chain scission yield, and q0 is the crosslinking yield; u2,0 can be calculated as the ratio of average molecular weight to average weight of a monomer unit. Degradation activation energy (E) of a crosslinked polymer can be calculated by using the following FreemaneCarroll equation Eq.(2.5) and the data obtained from derivatives of dynamic thermograms of an UV aged polymer.

(2.1)

where GHI, Global Horizontal Irradiance (W/m2), Tilt, the angle between the sun and the latitude-tilt POA (plane of array), and Zenith, the angle between the sun and the zenith.

 exp½3:473  0:0594  WS

AND

(2.3)

where aG, aC are the expansion coefficients of the glass (superstrate) and the cell respectively, D is the cell width, and C is the cell center to center distance. Equation 2.3 can also be applied to any material interface among substrate, superstrate, and encapsulant.

2.2.4.3 UV Crosslinking and Degradation Reaction The Charlesby-Pinner equation [12] has been widely used for simultaneous determination of the

Dlogðd%=dTÞ E Dð1=TÞ ¼n $ Dlogð1  cÞ 2:3R Dlogð1  cÞ

(2.5)

where T is the temperature in Kelvin, n is the order of reaction, R is the gas constant, and c is the conversion ratio and is equal to (m0  m)/m0, where m0 is the initial mass and m is the mass at any time.

2.2.4.4 Physical Aging Process The polymeric materials will experience a continuous change in physical and mechanical properties, such as specific volume, compliance, modulus, and impact strength as a function of aging time since the material requires a finite time to achieve thermodynamic equilibrium [14,15]. Therefore, the change in the above properties of polymeric materials used in PV modules due to physical aging can cause electrical and fire safety concerns for their long-term usage. The physical aging or enthalpic relaxation of amorphous polymeric materials is accompanied with the development of endothermic peaks on the glass transition with increasing aging time, and they can be measured by DSC [16]. The extent of enthalpic relaxation F(t) with aging time can be described by the WilliamseWatts exponential function [17], i h FðtÞ ¼ exp  ðt=s0 Þb ¼ 1  ðDHt =DHN Þ (2.6) where DHN ¼ DCp(Tg  Ta) is the maximum extent of physical aging, and s0 is the characteristic

30

D URABILITY AND R ELIABILITY

OF

P OLYMERS

relaxation time dependent both on the aging temperature Ta and the structure of the glass through the fictive temperature TF, and the structural parameter, x, and it can be defined by, s0 ¼ A exp½x=RTa þ ð1  xÞ=RTF DH



(2.7)

where A is the preexponential factor and DH* is the activation enthalpy for enthalpic relaxation. Recently, Schweizer [18,19] has developed the formulation of predictive theories of physical aging, the influence of deformation on the alpha relaxation process, and rate-dependent nonlinear mechanical properties of thermoplastics.

2.2.4.5 Moisture Ingress Process During long-term exposure of PV modules to environmental stress, the ingress of water into the module is correlated with decreased performance or increased failure rates [9,20]. Water can weaken interfacial adhesive bonds, resulting in delamination and increased ingress paths, consequent loss of passivation, electrochemical corrosion, and ultimately device failure. The water vapor transmission rate (WVTR) for a homogeneous material can be defined as WVTR ¼

P d

(2.8)

where d is the thickness of the film, and P is the permeability which is the product of the diffusivity (D) and solubility (S), P ¼ DS. The permeability often follows an Arrhenius dependence on temperature (T), P ¼ P0 eEp kT

(2.9)

where Ep is the activation energy for permeation and k is the Boltzmann constant. If it is assumed that the diffusivity obeys Fick’s law, then the transient WVTR as a function of time can be described as [21]: "



N X DCs 1þ2 ð1Þn e WVTRðtÞ ¼ d n¼1

 Dn2 p2 t d2

#

(2.10)

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

where Cs is the saturation concentration and t is the time.

2.2.4.6 Polymers/Materials Flammability The flammability of polymers is mainly evaluated through ignitability, flame spread, and heat release. Numerous tests, such as UL94, limited oxygen index (LOI), oxygen bomb calorimetry, and cone calorimetry, have been developed and performed either in a laboratory environment or during industrial manufacturing on final products. UL 94 has been generally applied for rating the flammability of polymers, but it does not provide a quantitative scale to compare their degree of flammability. In addition to complying with electrical safety requirements, certain polymeric products must, in order to meet certain safety regulations, meet largescale flammability performance requirements. Some popular tests for the polymer materials flammability are currently applied and are summarized with their sample dimensions and flammability information/ results or ratings as shown in Fig. 2.5. Each method offers a different perspective of polymer flammability. For example, UL 94 is an ignitability test that provides a ranking of materials ranging from lowest to highest performance: HB, V2, V1, and V0; the LOI determines the minimum concentration of oxygen that will support downward burning (candle-like combustion) of a vertically mounted test specimen; the cone calorimeter provides a broad range of ignition and combustion properties in a well-ventilated early fire-growth stage; and the oxygen bomb calorimeter measures the potential heat or calorific value under high pressure and a 100% oxygen environment. UL 94 has been generally applied for rating the flammability of polymers, but it does not provide a quantitative scale to compare their degree of flammability. However, this method has some significant drawbacks: (1) it is evaluated by visual observation, (2) it only measures the burning rate and drips, (3) the rating result depends on sample dimensions especially on thickness, (4) inconsistent results depend on the operators, and (5) scientific data/factors for rating are lacking. Therefore, we need a more rigorous testing tool for rating the degree of polymer flammability. Recently, the microscale combustion calorimeter (MCC) was successfully developed by Dr. Richard

2: OVERVIEW: D URABILITY AND R ELIABILITY

OF

C OMMON PV M ODULE

AND

P OLYMERS /M ATERIALS

31

Figure 2.5 Summary of the polymer flammability tests from macro- to microscale techniques.

Lyon of United States Federal Aviation Administration (FAA) [22,23], and has become available commercially for use by research organizations, universities, and industry. The MCC has been successfully applied to assess flammability of polymers using small milligram-sized samples from a pellet, film or powder; the sample size ranging from 2 to 10 mg [22e24]. The methodology and apparatus for MCC measurements have been standardized as ASTM D7309 [25]. In this standard, the sample is thermally decomposed and oxidized in a furnace, and the combustion products are analyzed, using oxygen consumption calorimetry, to determine the heat generated. The flammability index (Findex) was postulated by using three important flammability attributes from the MCC results: (1) the onset of combustion temperature (Ti), (2) maximum heat release rate ((HRR)m) or heat release capacity (hc ), and (3) heat of combustion (HOC). Each factor is determined by dividing the measured value by a reference value. The reference values allow for scaling of each of the contributing factors. Thus, changing the reference values will not qualitatively change the Findex. The reference values are selected based upon statistical results developed for UL 94 rated materials and average border values between high (HB) and low (V0) flammability materials with reference values of (1) hc: 200 J/goK, (2) HOC: 30 kJ/g, and (3) Ti: 300 C, used for evaluating the degree of flammability of polymers. Then, an empirical equation of Findex could rationally be established as in Eq.

(2.11) (please note that hc which is independent of heating rate was used instead of (HRR)m) to estimate the degree of flammability as proposed by Yang et al. [26,27]:  Findex ¼

k1 *hc 200



  k2 *ðHOCÞ k3 *ð300Þ 30 Ti (2.11)

where ki is the scaling parameter of each major factor. The above proposed equation indicates that the larger the Findex, the higher the flammability or risk of a flammability hazard.

2.2.5 Experimental Setup for the Durability and Reliability Study Accelerated Testing Equipment: The PV modules and their materials are characterized and evaluated for performance and safety properties prior to exposure and after accelerated aging/weathering as a function of exposure time and/or solar radiant intensity. Accelerated aging/weathering conditions may include novel and traditional accelerated protocols shown in Table 2.3. The aging/weathering equipment and test were determined by considering the optimum conditions to simulate as on the field exposures for average and most severe locations. Prior to and after exposure, materials or their modules are characterized for inherent properties and evaluated for key performance and safety properties.

32

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Table 2.3 Accelerated Aging and Weathering Equipment Accelerated Aging/ Weathering

UV

Heat

Moisture

NIST SPHERE*

X

X

X

Fluorescence

X

X

X

Xenon

X

X

X

Metal halide

X

Thermal oven

X

NIST-SPHERE*: National Institute of Standards and Technology (NIST) of USA has developed such Simulated Photo-degradation via High Energy Radiant Exposure (SPHERE) [28,29]. The detailed descriptions and results of the SPHERE technology and its application for the PV studies are discussed in Chapter 7.

2.2.5.1 Accelerated Testing Conditions Long-term sequential test (LST) is an accelerated test sequence for PV modules that is designed to allow a more accurate approximation of real outdoor solar module long-term operation via a multivariable sequential approach. This sequence has been designed specifically for this purpose and allows both manufacturers and end users to increase their confidence in the PV modules quality. Table 2.4 outlines the weathering conditions of damp heat, UV, thermal cycle, and/or humidity freeze (HF) with certain time or cycles that were generally performed for most of PV modules and materials. The key challenge is about what is the correlation of the performance results between such accelerated weathering tests and actual field exposure remains and needs to be resolved in the future. Often these accelerated exposures do not induce similar degradation mechanisms or rates as real-world degradation.

2.3 Observed Issues for Durability/ Reliability of PV Materials on the Field 2.3.1 Major Types of Field Failures for PV Components 2.3.1.1 Corrosion Moisture permeation through the module backsheet or through edges of module laminates causes corrosion and increases leakage of currents. Corrosion attacks cell metallization in crystalline silicon modules and semiconductor layers in thin-film modules (Fig. 2.6), causing loss of electrical performance. Therefore, when materials are not properly specified and tested, or ineffective encapsulation schemes are used, significant corrosion and loss of module power can be expected [4,30].

Table 2.4 The Accelerated Weathering Conditions for Testing PV Modules and Materials Testing Source Damp heat (DH)

Conditions 

85 C, 85% for 1000 h or more

UV (sunlight)

7.5 kWh/m2@280e320 nm, 15 kWh/m2@320e400 nm for 50 days or longer

Thermal cycle

At 40 C then at 85 C for 6 h each for 200 cycles or more

Thermal cycle (TC) and humidity freeze (HF)

(TC: 50 cycles, 40 C, >90 C) and (HF: 40 C, 85 C/85% RH, 10 cycles, 24 h/cycle)

2: OVERVIEW: D URABILITY AND R ELIABILITY

OF

C OMMON PV M ODULE

AND

P OLYMERS /M ATERIALS

33

Figure 2.6 Busbar corrosion; corrosion of cell interconnects or soldering joint.

2.3.1.2 Solder Bond Failures As solder bonds “age” due to continuous thermal cycling outdoors, expansion and contraction cause the solder to fatigue, become more brittle, and to disassociate into large grains of tin and lead. These phenomena result in a propensity for the solder bonds to crack, becoming more resistive with age [31].

2.3.1.3 Cell or Interconnect Break This is a common failure mode, although redundant contact points plus “interconnected-busbars” allow the cell to continue functioning. Cell cracking can be caused by [32]: 1. thermal stress;

Module manufacturers have set out to develop a test sequence (e.g., dynamic mechanical load cycles [DMLC]) that would evaluate the potential for cell cracking within PV modules (Fig. 2.7) [33].

2.3.1.4 Encapsulant Discoloration Cross-linked EVA sheet is a common encapsulant for PV modules. In very hot regions with high solar radiation, the encapsulant often darkens over time (Fig. 2.8), reducing the amount of solar energy converted to electricity. The cause and extent of discoloration have been reported as due to changes in the polymer and effects of the stabilizer and peroxide cross-linking additive system. Research has now led to improved understanding of the EVA polymer and to new formulations that are more UV tolerant.

2. hail; or 3. damage during processing and assembly, resulting in “latent cracks,” which are not detectable on manufacturing inspection, but appear sometime later.

2.3.1.5 Junction Box Failures Junction Box related failures could occur due to a number of reasons. A quite frequently detected defect

Figure 2.7 Crack of solar cells. Leftdwith the Optical Microscope; Rightdwith Electro-Luminescene (EL) Detection Test.

34

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 2.8 Examples of EVA discoloration.

Figure 2.9 Leftddetachment of a J-Box; Rightdjunction box degradation/burning.

is bad adhesion of the junction box on the PVF backsheet (Fig. 2.9, left) with the consequent risk of detachment when opening, loss in insulation resistance, and water penetration. Others like poor workmanship, degradation of enclosure materials (Fig. 2.9, right), corrosion of connectors, open circuit due to backsheet delamination or creep, and short circuit results from insulation failure or through failure of a bypass diode.

2.3.1.6 Delamination Module delamination, resulting from loss of adhesion between the encapsulant on other module layers, is also a failure mechanism that needs to be addressed in order for manufacturers to achieve 30year product lifetimes. From an industry-wide perspective, delamination has occurred in a small percentage of modules manufactured since the mid1980s. However, it has occurred to varying degrees in modules from all manufacturers, and, because the causes for the failure mechanism are not well understood, it is a continual quality control issue for manufacturers. Most of the delamination observed in the field has occurred at the interface between the

encapsulant and the front surface of the solar cells in the module (Fig. 2.10). A common observation has been that delamination is more frequent and more severe in hot and humid climates, sometimes occurring after less than 5 years of exposure. Delamination first causes a performance loss due to optical decoupling of the encapsulant from the cells. Of greater concern from a module lifetime perspective is the likelihood that the void resulting from the delamination will provide a preferential location for moisture accumulation, greatly increasing the possibility of corrosion failures in metallic contacts. Unfortunately, typical accelerated-aging tests have not been effective in accelerating the mechanisms responsible for delamination, making laboratory investigation of the phenomenon more difficult [29]. It is further noted that delamination and cracking of the multilayer backsheets often occurred [34].

2.3.1.7 Hot Spots Hot-spot heating occurs in a module when its operating current exceeds the reduced short-circuit current (Isc) of a shadowed or faulty cell or group of cells. When such a condition occurs, the affected

2: OVERVIEW: D URABILITY AND R ELIABILITY

OF

C OMMON PV M ODULE

AND

P OLYMERS /M ATERIALS

35

Figure 2.10 Left: splitting delamination of the backsheet; Right: encapsulant delamination.

cell or group of cells is forced into reverse bias and must dissipate power. If the power dissipation is high enough or localized enough, the reverse biased cell can overheat resulting in melting of solder and/or silicon and deterioration of the encapsulant and backsheet (Fig. 2.11). With more BIPV and higher cell efficiency, more attention should be paid to the hot spot damage [35].

2.3.1.8 Mechanical Damage/Glass Breakage The useful life of photovoltaic modules may depend on their ability to withstand periodic exposure to high wind forces, cyclic loads induced by specific site conditions or shipment methods, high loads caused by accumulated snow and ice on the module surface, and twisting deflections caused by mounting to non-planar surfaces or structures. The effects on the module may be physical or electrical, or both. Most importantly, the effects may compromise the safety of the module, particularly in high voltage applications, or where the public may be exposed to broken glass or other debris. Shattering of

the top glass surface can occur due to vandalism, thermal stress, handling, wind, or hail [32]. Failure of or misuse of support structures can also lead to glass breakage (Fig. 2.12). It is important to note that Corning Incorporated has recently produced photovoltaic (PV) Willow glass substrates with stronger impact, higher efficiency, and lighter weight thin-film photovoltaic modules. Recently, the US government’s National Renewable Energy Laboratory (NREL) has built flexible solar cells out of Corning’s Willow Glass. These new solar cells are strong enough that they could eventually replace roofing shingles, which would significantly shrink the biggest barrier to mass adoption of solar power: the cost of installation. Willow Glass’s combination of flexibility, transparency, and heat resistance does make it a very good option for the creation of cadmium telluride (CdTe) solar cells, however. The CdTe photovoltaic (PV) cells are the only thin-film PV technology that is cheaper than crystalline silicondbut until now, there has not been a transparent substrate that also has the thermal resistance to withstand manufacturing [36].

Figure 2.11 Heat dissipated in a shaded cell caused the module to crack.

36

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 2.12 Snow load can deform the frame and break the glass.

2.3.1.9 Defective Bypass Diodes Bypass diodes protect the module from overheating when the module is shadedda condition hard to avoid in many installations. Axial diodes potted in the junction box are prone to failure because they cannot dissipate heat. Diode failure is caused by overrating of the diode (diode manufacturer issue), and/or by inappropriate electrical configuration and short circuit current of the module (module manufacturer issue). It must be properly rated for the highest temperature that can occur, or placed on a heatsink outside of the box, depending on ambient temperature, air flow around the modules, etc. The problem is minimized if junction temperatures are kept below 128 C [32].

2.3.1.10 Arcing [33] In a high voltage DC system any open circuit can lead to formation of an arc. Arcs have been observed in most of the components utilized in PV systems. Modules are not immune although most of the wiring within a module has a degree of redundancy. Most

module arcing occurs where there are single point connections like input wiring in the junction box (Fig. 2.13, left). Input wire failures are the most likely since they only require a failure of one connection and are often a manual process. A second type of module arcing problem can occur where there are parallel connections like the two interconnects on a cell. In the case not only must both of the interconnect ribbons fail (Fig. 2.13, right), but the bypass diode must also malfunction or not be correctly connected for an arc to occur.

2.3.1.11 Inverter Inverters have historically been the leading cause of PV system failures [37]. The useful life of a central inverter typically does not exceed 10 years, and the cost to maintain and eventually replace a central inverter once or even twice during a PV system’s lifetime drives up system costs with every truck roll. There are a variety of accelerated tests required at the assembly level to confirm that failure mechanisms were not missed during the design review

Figure 2.13 Left: J-box burned due to bad DC connection; right: arc across cell to cell interconnection.

2: OVERVIEW: D URABILITY AND R ELIABILITY

OF

C OMMON PV M ODULE

process. Each test applies a set of stresses to induce specific failure mechanisms. Common tests and typical failure mechanisms they may identify are:
Highly Accelerated Life Test (HALT):
design margin and
confirmation of soft shutdown at operation limits;
Thermal Cycle (TC):
solder joint fatigue and
PCB via fatigue;
Damp Heat
conductive anodic filament (internal PCB short) and
cracked ceramic capacitor;
Humidity Freeze (HF)
interfacial stresses leading to delamination and
dendrite growth; and
High Temperature Operating Bias (HTOB)
E-cap electrolyte vaporization and
ceramic capacitor oxygen vacancy migration.

AND

P OLYMERS /M ATERIALS

37

surfaces of the connector, leading to a build-up of wear debris, which can result in eventual connector failure. Certain metallic platings for copper, the conductor of choice, are less prone to fretting and corrosion than othersdideally gold would always be used but cost dictates alternatives are required for PV systems. Tin is often used as a plating metal being resistant to corrosion and of low cost. However, tin has two major drawbacks (1) tin rapidly forms an insulating oxide in air which must be penetrated by the surfaces of a connector when mated and (2) tin is prone to fretting wear. Alternative platings such as silver and palladium are also used but these are more expensive. The working lifetime of the connectors is limited not only by the effects of fretting. Poor fabrication and assembly at installation are the most common causes of failure. For example, the failure of the environmental seal will dramatically shorten a connector’s life [38].

2.3.2 Polymer Materials Related Problems Table 2.5 provides a summary for failure modes or degradation of PV materials for substrates by observation (⇧: increase, ⇩ decrease).

2.3.1.12 PV Connector

2.3.2.1 EVA Discoloration

PV connectors within roofs and facades expand and contract on a daily basis, which can lead to small amplitude movement at the interface surface of a mated connectordthis is commonly known as fretting. In time, the fretting process wears the

The discoloration of EVA is primarily due to the formation of polyconjugated C]C bonds (polyenes) generated by multistep deacetylation, and from the presence of a, b-unsaturated carbonyl groups, which can be UV-excited leading to photodegradation of

Table 2.5 Observed Failure Modes of PV Materials Encapsulant

Substrate (Backsheet)

Yellowing

⇧

Delamination

⇧

Hazing

⇧

Discoloration (PVF)

⇧

Crack

⇧

Crack

⇧

Transparency

⇩

Embrittlement (PET, PVDF)

⇧

Creep and Flow

⇧

Creep

⇧

Interfacial Adhesion (glass, back sheet)

⇩

Barrier (Vapor, Oxygen)

⇩

Water Ingress

⇧

38

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

-----(CH2CH2)n(CH2CH)m(CH2CH2)n---------(CH2CH2)n(CH=CH)m(CH2CH2)n----O

(Norrish ΙI) (Deacetylation)

O=C-----CH3 (Norrish I)

(Polyenes) + m CH3COH

UV, T

O (Acetic acid)

(Norrish ΙII) -----(CH2CH2)n(CH2CH)m(CH2CH2)n-----

-----(CH2CH2)n(CH2C)m(CH2CH2)n-----

O .

O (Ketone)

+

+ m CH3C--H

O= C---CH3

O (Aldehyde)

RH

CH3C---H

+R.

CH4, CO2, and/or CO

O

H C H

H C H

H C C H

H C C H

H C C H

H C H

UV, T

O2

H C H

(Polyene oxidation)

H C H

C H

C H

H C H

OOH H C C H

H C

H C C H

C H

- H2O

H C H H C H

- H2O

H C H

H C C H

O H C H

H C C H

C

OOH H C H

H C

H C

H C C

H C C H

C H

O

C C H

H C

C H

H C

C H

H C

H C H

(a,β-unsaturated carbonyls)

Figure 2.14 Most significant degradation mechanisms of ethylene vinyl acetate copolymer are induced from UV exposure and temperature (composite from the literature) [39].

2: OVERVIEW: D URABILITY AND R ELIABILITY

OF

C OMMON PV M ODULE

AND

P OLYMERS /M ATERIALS

39

polyolefins (Fig. 2.14). Furthermore, acetic acid produced by thermal and photothermal decomposition exhibits an autocatalytic effect on the yellowing or browning of EVA [39]. Hydrolysis of vinyl-acetate monomers resulted in generating acetic acid that can accelerate the corrosion of electrical interconnects. Hydrolysis of EVA and other polymers is accelerated by temperature, oxidation, irradiance, manufacturing processes, and the combined effect of these stressors [40]. There is a multifactor response to the degradation of EVA. Recently, scientists in Mitsui Chemical were using InfraRed (IR) and Hot Water Extraction Method (HWEM) to evaluate the amount of free acetic acid desorbed in the EVA encapsulant. They proposed that the desorbed acetic acid in EVA was useful to be used as an indicator for module degradation [41,42]. In this study, the free acetic acid was evaluated by a Pyrolysis GC-MS (Frontier Lab) on full size PV modules [43].

2.3.2.3 Embrittlement

2.3.2.2 Creep

2.3.2.4 Interfacial Adhesion

During operation, photovoltaic (PV) modules are subject to elevated temperature conditions. This may enable phase transformation or viscoelastic flow to occur within polymeric components. This flow or creep of critical PV polymeric components can result in a risk of shock, fire, or mechanical hazards.

The sample for the PV component interface test may be prepared as shown in Fig. 2.15. Teflon film was used for separating the partial interface for a peel test. The interfacial performance and properties of the interface samples were evaluated before and after exposure.

EVA suffers from having both glass and melting phase transitions at temperatures experienced under environmental exposure. These transitions cause EVA to embrittle at low temperatures (w-15 C) and to be very soft at high temperatures (>40 C). For many environments a temperature of 15 C is often reached, making cells in EVA-based modules significantly more susceptible to breakage from sudden impacts and rapid flexing [44]. For the backsheet, fluoropolymers such as polyvinyl fluoride (PVF) and polyvinylidene fluoride (PVDF) act as a protection against weathering influences, while polyesters such as polyethylene terephthalate (PET) provide mechanical strength. In several publications, embrittlement of PVDF and PET has been described. The embrittlement can be attributed to global aging and chemical degradation of the materials [45].

Encapsulant/Superstrate Substrate (Backsheet) Teflon

Encapsulant Superstrate

Encapsulant/Substrate Substrate (Backsheet) Teflon

Encapsulant Superstrate

Encapsulant/PV Cell Substrate (Backsheet) Teflon

Encapsulant

PV Cell

Superstrate

Figure 2.15 The sample preparation designed and used for the interfacial adhesion study.

40

D URABILITY AND R ELIABILITY

OF

P OLYMERS

The adhesion strength between the glass and EVA of the coupons was measured with an Instron test system using a 90 peel test. As mentioned before, the specific coupons were fabricated using different backsheets with and without edge sealing tape. The results [46] showed that the adhesion strength of UVaged samples dramatically dropped from 15 to 1 N/ mm2 at 1000 h, and then, it remained constant. For damp heat-aged samples, it gradually decreased from 15 to 10 N/mm2 and 3 N/mm2 at 1000 h and 2000 h, respectively, and then, it remained stable at around 3 N/mm2. The results demonstrated that UV has a greater aging effect than damp heat on the adhesion strength between glass and EVA. However, the peeling test results did not show any significant difference in EVA adhesion between PVF/ PET/PVT and PVF/PET/EVA substrates. The coupons sealed with the edge sealing tape showed better adhesion strength under damp heat conditions only at 1000 h. This might be because the edge sealing tape was saturated with water vapor at 1000 h, causing it to lose its function of preventing water ingress for longer exposure times. The mechanisms of interfacial adhesion loss were different between damp heat and UV conditions. Damp heat exposure causing adhesion loss was associated with continuous moisture diffusion into the interfacial sites while the UV caused adhesion loss was related to the irradiation energy of UV being much larger than that for polymer bond dissociation [47].

2.3.2.5 Thermal Expansion PV modules have a multilayer structure. Different thermal expansions of the materials used in a PV module (glass, solar cell, interconnects, encapsulant and backsheet) can result in over-stressing and cracking. To avoid this, the material used has to be a low modulus, elastomeric material [48].

2.3.2.6 Sealant Degradation An elastomeric or polymeric module edge seal may become brittle from UV and thermal degradation leading to moisture ingress and resultant electrical failure.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

2.4 PV Materials Performance and Service Life Prediction: Challenges and Future Perspective 2.4.1 Development of Improved or New Materials Recently, some improved encapsulant materials have been developed. Arkema Co. and Wacker Chemie introduced a highly transparent nanostructured thermoplastic polymerdApolhya® Solar and a thermoplastic-organo-silicon elastomerd Tectosil® for use without a curing process, respectively. DuPont recently developed PV5200 series encapsulant sheets based on polyvinylbutyral (PVB) polymer technology and PV5400 ionomers based encapsulants, and Dow ENGAGE PV Polyolefin Elastomers (POEs) especially for the need of improved power efficiency in the growing thin-film module technology market. They all claim that the above encapsulant materials show excellent glass adhesion, proven safety glass performance with high visible light transmission, increased protection from moisture, and improved power retention. Corning has also developed high impact and much lighter Willow and Gorilla glass to solve the breakage problem.

2.4.2 Qualification Test Methods Reliability testing is a set of well-defined accelerated stress testsdirradiation, environmental, mechanical, and electricaldwith strict passefail criteria based on functionality/performance, safety/insulation, and visual requirements. Qualification testing does not, as anticipated, identify all the possible lifetime/reliability issues that would be encountered in the field; however, it does identify the major/ catastrophic design quality issues that would initially occur in the field (except that it did not identify the polyamide failure of backsheets). The type, extent, limits, and sequence of the accelerated stress tests of the qualification standards have been stipulated with two goals in mind: one, accelerate the same failure mechanisms observed in the field but without introducing other unknown failures that do not occur in the field; and two, induce these failure mechanisms in a reasonably short period of time (60e90 days) to reduce testing time and cost.

2: OVERVIEW: D URABILITY AND R ELIABILITY

OF

C OMMON PV M ODULE

Currently the global PV qualification test standards can be separated into: a. US Certifications
UL 1703dStandard for Flat-Plate Photovoltaic Modules and Panels,
UL 1741dStandard for Inverters, Converters, Controllers and Interconnection System Equipment for Use with Distributed Energy Resources,
UL 790dStandard for Standard Test Methods for Fire Tests of Roof Coverings,
UL 8703dPV Concentrator Requirements, and
Building Integrated Photovoltaic (BIPV) Module Requirements (AC 365); b. IEC Standards, European Certifications
IEC 61730, Photovoltaic Module Safety Qualification,
Part 1: Requirements for Construction and
Part 2: Requirements for Testing,
IEC 61215, Crystalline Silicon Terrestrial Photovoltaic (PV) Modules Design Qualification and Type Approval,
IEC 61646, Thin-film Terrestrial Photovoltaic (PV) ModulesdDesign Qualification and Type Approval,
IEC 61218, Concentrator Photovoltaic (CPV) Modules and AssembliesdDesign Qualification and Type Approval, and

AND

P OLYMERS /M ATERIALS

41


IEC 60904-X, PV DevicesdMeasurement Procedures and Requirements.

2.4.3 Correlation Between Outdoor Performance and Laboratory Accelerated Testing In order to study the correlation between lab accelerated test and actual field performance of PV modules/panels, we may fabricate a mini-module with two by two cells as shown in Fig. 2.16. Such a mini-module can be easily tested in the lab. Correlation functions of both materials degradation and failure modes between lab and field can be developed and used for predicting service life. Field experience is a critical element for identifying real design failure modes. The stress levels must accelerate the same failure mechanism observed in the field. Current qualification tests are based on known field failure mechanisms. They likely will not identify failure mechanisms that [37]:
appear after longer-term exposure outdoors, socalled end of life failures,
can be caused by combinations of stresses, and
occur with new technologies for which field data does not exist. Failure analysis must be done on each type of filed failure to determine what combination of stresses caused the failure, so a test program can be devised to ensure that the new products do not have the same weaknesses.

Figure 2.16 The experimental design for studying the correlation between lab accelerated and actual field performance.

42

D URABILITY AND R ELIABILITY

OF

P OLYMERS

2.4.4 Correlation Between the Degradation of PV Materials and Module Failure Modes Recent results from visual inspections and studies on used PV modules indicated that most of defects, degradations, and/or failures arising from materialrelated issues are:
discoloration of encapsulant and substrate,
oxidation or browning of cell electrical grid,
encapsulant delamination,
substrate delamination,
junction box detachment, and
glass cracking. Therefore, it is important to investigate how properties change and/or degrade in polymeric materials used in PV modules and understand the correlation among materials degradation and failures of PV module and system performance in the field [48] and lab as shown in Fig. 2.16. Recent field surveys of backsheet degradation suggest that the location of a module in the field plays a key role in the level of degradation of backsheets since rear-side irradiance is not uniform in a field [49].

2.4.5 PV Service Life Prediction In order to predict the service life of PV module, we may first identify the major failure mode(s) via a material degradation approach from durability and reliability studies in the field [43], then the correlation factor for results in the field and lab. The statistical and empirical approach may be needed to precisely predict the service life of PV modules as a function of weathering conditions, such as, Temperature (T), UV Intensity (I), Relative Humidity (RH), and geographical location, as shown in the following Eq. (2.12): FðT; I; RH; tÞ ¼ c1 F1 þ c2 F2 þ c3 F3 þ c12 F1 *F2 þ c13 F1 *F3 þ c23 F2 *F3 þ c123 F1 *F2 *F3 þ c0 F0 (2.12)

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

where 1. UV Exposure: F1 ¼ FUV (I, t), 2. Temperature: F2 ¼ FT (t), 3. Relative Humidity (RH): F3 ¼ FRH (%, t), 4. F0 from other internal and external stress factors, and 5. ci is the contribution parameter to the function of each weathering exposure time. However, the above proposal will face several challenges: first, it takes a very long time and tedious study to develop such an empirical equation; secondly, we need to find out the aging process mechanism and correlation factor between accelerated lab and actual field-tested results; thirdly, to identify the major failure mode(s) of the material(s) for the failure and/or safety of PV module. Recent data driven methods have approached lifetime and degradation of PV modules with a new approach. Data driven methods focus on collecting large data sets of both various accelerated and realworld exposures of PV materials and modules (stressors), and then developing predictive data driven models of performance responses. Detailed analysis of degradation mechanisms allows for network models that relate stress to degradation mechanisms to performance responses. By including both real-world exposed materials and modules, degradation mechanisms that occur (rate or mechanism type) differently in accelerated exposures can be identified. This approach is key to move beyond qualification testing into accurate lifetime prediction [4,5,50]. The advance in data driven methods to understanding materials and PV module degradation is due to the development of new ways to handle and query large stores of data such as nonrelation data warehouses [51].

2: OVERVIEW: D URABILITY AND R ELIABILITY

OF

C OMMON PV M ODULE

AND

P OLYMERS /M ATERIALS

43

Appendix I Table 1 Current Standards for Polymeric Materials Used in PV Modules Property

Requirement

Required Test(s)

UL1703

IEC61730

Applications

5VA (enclosure)

UL94 HB, V2, V1, V0, 5VA IEC 60695-11-20

Section 7.1, 7.2

Section 5.2, 5.3

Enclosure, support

UV and Water resistance

f1 or f2 (limited)

ANSI/UL 746C

Section 7.1, 7.2

Section 5.2, 5.3, 5.4

Enclosure, support, outer surface

Hot Wire Ignition (HWI)

30 (IEC 61730)

UL 746A IEC 60695-11-20

Section 7.1

Section 5.2

Enclosure

High Current Arc Ignition (HAI) for given flammability classification

See table

UL 746A IEC 60695-11-20

Section 7.2

Section 5.3

Support or insulation

Comparative tracking Index (CTI)

PLC 2 or 250V

UL 746A IEC 60112

Section 7.2

Section 5.3

Support or insulation

Relative Thermal Index (RTI)

>90 C, >MOT þ 20 C

UL 746 B IEC 60216-5

Section 7.3

Section 5.4

Outer surface, all polymers

Flame Spread Index (FSI)

100

ASTM E 162-02a

Section 7.4

Section 5.4

Outer surface

Thickness and appropriate material

e

IEC 61140

Section 7.5

Section 5.5

Barriers

Flammability

Acknowledgments I would like to thank Ethan Wang of Underwriter Laboratories (UL) for his many valuable discussions and cooperation while I was working in UL.

References [1] Data was taken from US Energy Information Administration (EIA), 2017. [2] IEC 61730, Photovoltaic (PV) Module Safety Qualification - Part 2: Requirements for Testing. International Electrotechnical Commission, Geneva, Switzerland, 2016.

[3] UL 1703, Standard for Flat-Plate Photovoltaic Modules and Panels, Second Edition, Underwriters Laboratories Inc., 333 Pfingsten Road, Northbrook, IL 60062, 2013. [4] L.S. Bruckman, N.R. Wheeler, J. Ma, E. Wang, C.K. Wang, I. Chou, J. Sun, H. Roger, French, Statistical and domain analytics applied to PV module lifetime and degradation science, IEEE Access 1 (2013) 384e403. [5] R.H. French, R. Podgornik, T.J. Peshek, L.S. Bruckman, Y. Xu, N.R. Wheeler, A. Gok, Y. Hu, M.A. Hossain, D.A. Gordon, P. Zhao, Degradation science: mesoscopic evolution and temporal analytics of photovoltaic energy

44

[6] [7]

[8]

[9] [10]

[11] [12] [13] [14]

[15] [16] [17] [18] [19] [20]

[21] [22] [23]

[24]

D URABILITY AND R ELIABILITY

OF

P OLYMERS

materials, Current Opinion in Solid State and Materials Science 19 (4) (2015) 212e226. J.H. Wohlgemuth, NCPC and Solar Program Review Meeting, 2003. A. Realini, Mean Time before Failure of Photovoltaic Modules, SUPSI, DACD, LEEE-TISO, June 2003. J. Pern, Module encapsulation materials, processing and testing, in: NCPV/NREL, Presentation to APP International PV Reliability Workshop, December 2008. NREL-CP-520-37390, Control of Moisture Ingress into PV Modules, M. D. Kempe, February 2005. D.L. King, W.E. Boyson, J.A. Kratochvil, Photovoltaic Array Performance Model, Sandia National Labs, 2004. SAND2004-3535. Solar Engineering of Thermal Processes, second ed., John Wiley & Sons, Inc., 1991, p. 25. A. Charlesby, Atomic Radiation and Polymers, WNT, Warsaw, 1962. M. Copuroglu, M. Sen, Polym, Advanced Metals Technology 16 (2005) 61e66. L.C.E. Struik, Physical Aging of Amorphous Polymers and Other Materials, Elsevier, New York and Amsterdam, 1978. R.N. Haward, The Physics of Glassy Polymers, Applied Sci. Pub. Ltd., London, 1973. J.N. Hay, Pure and Applied Chemistry V 67 (11) (1995) 1855e1858. G. Williams, D.C. Watts, Transactions of the Faraday Society 67 (1994) 1323. K.S. Schweizer, E.J. Saltzman, The Journal of Chemical Physics 121 (2004) 1984. K. Chen, K.S. Schweizer, The Journal of Chemical Physics 126 (2007) 14904. G.J. Jorgensen, et al., Solar Energy Materials & Solar Cells, vol. 90, Elsevier, New York and Amsterdam, 2006, pp. 2739e2775. J. Crank, The Mathematics of Diffusion, Clarendon Press, Oxfod, 1975, p. 50. R.E. Lyon, R.N. Walters, Micro-scale Combustion Calorimeter, U.S. Patent 5,981,290, 1999. R.E. Lyon, Heat Release Rate Calorimeter for Milligram Samples, U.S. Patent 6,464,391 B2, 2002. R.E. Lyon, R.N. Walters, Principles and Practice of Micro-Scale Combustion Calorimetry, DOT/ FAA/TC-12/53, U. S. Department of Transportation, Federal Aviation Administration,

AND

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32] [33] [34]

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Office of Aviation Research, Washington, DC, 2013, 20591. ASTM D7309 e Standard Test Method for Determining Flammability Characteristics of Plastics and Other Solid Materials Using MicroScale Combustion Calorimetry, ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA, 19428e2959, 2018. H.E. Yang, P. Gandhi, T. Lackhouse, Micro-scale evaluation of flammability for cable materials, in: Proceedings of 62th International Wire & Cable Symposium, Dec, 2013. H.E. Yang, F. Schall, Micro-scale study on the flammability of polymers, in: Proceeding of 2015 BCC Flame Conference, May 17e20, 2015, 6 pages. J. Chin, E. Byrd, N. Embree, J. Garver, B. Dickens, Accelerated UV weathering device based on integrating sphere technology, Review of Scientific Instruments 75 (11) (2004). X. Gu, D. Stanley, W.E. Byrd, B. Dickens, I. Vaca-Trigo, W.Q. Meeker, T. Nguyen, J.W. Chin, J.W. Martin, Chapter 1 e linking accelerated laboratory test with outdoor performance results for a model epoxy system, in: J.W. Martin, R.A. Ryntz, J. Chin, R.A. Dickie (Eds.), Service Life Prediction of Polymeric Materials e Global Perspectives, Springer, 2009, pp. 3e28. M.A. Quintana, D.L. King, T.J. McMahon, C.R. Osterwald, Commonly observed degradation in field-aged photovoltaic modules, in: Photovoltaic Specialists Conference, 2002, pp. 1436e1439. Conference Record of the 29th IEEE, 19e24 (May 2002). D.L. King, M.A. Quintana, J.A. Kratochvil, D.E. Ellibee, B.R. Hansen, Photovoltaic Module Performance and Durability Following Long-Term Field Exposure, Sandia National Laboratories, 2000. http://www.pveducation.org/pvcdrom/modules/ degradation-and-failure-modes. J. Wohlgemuth, Reliability of PV System, BP Solar, 2009. C.-C. Lin, Y. Lyu, D.L. Hunston, J.H. Kim, K.-T. Wan, D.L. Stanley, X. Gu, Cracking and delamination behaviors of photovoltaic backsheet after accelerated laboratory weathering, Proceedings of SPIE 9563 (2015) 956304.

2: OVERVIEW: D URABILITY AND R ELIABILITY

OF

C OMMON PV M ODULE

[35] J. Wohlgemuth, D.W. Cunningham, A. Nguyen, Failure Modes of Crystalline Si Modules, BP Solar, 2010. [36] K.L. Chopra, P.D. Paulson, V. Dutta, Thin-film solar cells: an overview, in: Issue 2 e 3 of “Progress in Thin Film Solar Cells”, vol. 12, May 2004. [37] System Reliability for Utility PV Inverters, Ron Vidano, 2014. www.nrel.gov/pv/assets/pdfs/ 2014_pvmrw_37_vidano.pdf. [38] B.B. Yang, N. Robert Sorensen, P.D. Burton, J.M. Taylor, A.C. Kilgo, D.G. Robinson, J.E. Granata, Reliability Model Development for Photovoltaic Connector Lifetime Prediction Capabilities, 2014. http://energy.sandia.gov/ wp-content/gallery/uploads/Yang_SAND20134705C_PVSC391.pdf. [39] A.W. Czanderna, F.J. Pern, Encapsulation of PV Modules Using Ethylene Vinyl Acetate Copolymer as a Pottant: A Critical Review, NREL, 1995. [40] S.S. Hosseini, S. Taheri, A. Zadhoush, A. Mehrabani-Zeinabad, Hydrolytic degradation of poly(ethylene terephthalate, Journal of Applied Polymer Science 103 (4) (2007). [41] K. Ohshimizu, Estimation of Amount of Free Acetic Acid Desorbed in Eva Encapsulant with Infra-red Spectrums, Mitsui Chemicals, Inc., 2012. [42] T. Shioda, Amount of desorbed acetic acid in EVA during DH test and long-term outdoor exposure and its influence on module performances, in: NREL PV Module Reliability Workshop, 2012. [43] H.E. Yang, E. Wang, et al., Failure modes evaluation of PV module via materials degradation approach, Science Direct, Energy Procedia 33 (2013) 256e264. Elsevier. [44] M.D. Kempe, Rheological and Mechanical Considerations for Photovoltaic Encapsulants, NREL, 2005.

AND

P OLYMERS /M ATERIALS

45

[45] G. Oreski, G.M. Wallner, Aging mechanisms of polymeric films for PV encapsulation, in: Polymeric Materials for Solar Energy Applications, 2005. [46] M.D. Kempe, Evaluation of encapsulant materials for PV applications, Photovoltaics International 9 (2010) 170e176. [47] G. Oreski, G.M. Wallner, Damp Heat Induced Physical Aging of PV Encapsulation Materials, Polymer Competence Center Leoben GmbH, 2010. [48] E. Wang, C.-Y. Peng, C.-C. Tsai, I. Chou, C. Wang, Correlation between Material Degradation Behavior and PV Module Performance, in: PV Conference, 2014. [49] A. Fairbrother, M. Boyd, Y. Lyu, J. Avenet, P. Illich, Y. Wang, M. Kempe, B. Dougherty, L.S. Bruckman, X. Gu, Differential degradation patterns of photovoltaic backsheets at the array level, Solar Energy 163 (2018). [50] A. Gok, D.K. Ngendahimana, C.L. Fagerholm, R.H. French, J. Sun, L.S. Bruckman, Predictive models of poly(ethylene-terephthalate) film degradation under multi-factor accelerated weathering exposures, PLoS One 12 (2017), 0177614. [51] V.Y.G. Hu, P. Zhao, D. Gordon, N.R. Wheeler, M.A. Hossain, T.J. Peshek, L.S. Bruckman, G.Q. Zhang, R.H. French, A nonrelational data warehouse for the analysis of field and laboratory data from multiple heterogeneous photovoltaic test sites, IEEE Journal of Photovoltaics 7 (1) (2017) 230e236. https://doi.org/10.1109/ JPHOTOV.2016.2626919.

Further Reading [1] www.pveducation.org/pvcdrom, Chapter 7.

3 Degradation Science and Pathways in PV Systems Abdulkerim Gok 1, Devin A. Gordon 2,4, Menghong Wang 2,4, Roger H. French 3,4 and Laura S. Bruckman 3,4 1 Gebze Technical University, School of Engineering, Materials Science and Engineering, Gebze, Kocaeli, Turkey 2 Case Western Reserve University, Case School of Engineering, Macromolecular Science and Engineering, Cleveland OH, United States 3 Case Western Reserve University, Case School of Engineering, Materials Science and Engineering, Cleveland OH, United States 4 Case Western Reserve University, Case School of Engineering, SDLE Research Center, Cleveland OH, United States

Successful service life prediction models require good, time-parsed information about use conditions, knowledge of how environmental variables affect the rate of degradation, and good rate data under at least one set of accelerated conditions. There are various ways this information can be sliced and put into different models or approaches, but all the pieces must be present in some form. In the three-step approach, one can establish the kinetics for a material and use those kinetics to calculate service life from accelerated test data for a particular formulation using that material. The approach is not without risk since one cannot know when a formulation change is drastic enough to change the kinetics, but it may be the best that can be done with limited time and resources. J. E. Pickett [1].

3.1 Introduction The continued robust adoption of PV systems requires significant technological and market developments, not only in terms of increasing cell or module efficiency and decreasing manufacturing and levelized PV electricity costs [2e4], but also increasing reliability and sustainability for increasing long service lifetimes. In today’s PV market, modules are typically sold with 25 year power production warranties with less than 1% power loss each year (for example, at least 75% power production after

25 years) [5,6]. A problem arises because this warranty applies for modules deployed in all Ko¨ppene Geiger climatic zones [7,8] even though these PV modules face very different environmental exposure conditions such as extreme hot or cold, dry or humid, heavy snow loads or high-speed and gusty winds. Material selection and module design are therefore key parameters for climate sensitive degradation, and there is a strong research need in the PV community to address climate- and weather-induced reliability and degradation issues. Improving the lifetime performance of PV systems requires a better understanding of the active degradation mechanisms of the materials, components, and systems, under realworld exposure conditions. Reliability of engineering systems falls into three main regimes in the well-known bathtub curve as seen in Fig. 3.1: infant mortality, intrinsic random failures, and wear-out failures at the end of life [9,10]. Infant mortality failures usually occur due to manufacturing flaws, such as improper lamination, and during transportation and installation, such as glass or cell breakage due to mechanical impact, and affect modules immediately and dramatically in the beginning of their operation. Intrinsic failures are mostly related to defective metallization, cell, or string interconnections during module construction and failures due to junction box, bypass diode, and glass and frame breakage. The end-of-life wear-out failures are often associated with delamination and highly discolored and/or cracked encapsulants and backsheets due to strong weather degradation, and failures due to cracked solar cells and serious

Durability and Reliability of Polymers and Other Materials in Photovoltaic Modules. https://doi.org/10.1016/B978-0-12-811545-9.00003-3 Copyright © 2019 Abdulkerim Gok, Devin A. Gordon, Menghong Wang, Roger H. French & Laura S. Bruckman. Published by Elsevier Inc. All rights reserved.

47

48

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 3.1 Bathtub curve in reliability engineering of photovoltaic (PV) systems. Some degradation mechanisms observed in PV modules are included for the three regimes of failure.

metallization corrosion. Some examples of observed field degradation mechanisms in various environments can be found in these references [11e44]. Lifetime and degradation science (L&DS) is a robust scientific approach to reliability of long lifetime engineering applications based on developing a network structural equation model (netSEM) that encompasses the active degradation pathways and mechanisms present in these complex systems [45e61]. This new approach is based on linear response theory and a stressor (S) and response (R) framework and includes both predictive models and inferential mechanistic (M) pathway models. These models, based on structural equation modeling (SEM) [62,63], introduces functional forms beyond simple linear relationships, so that characteristic functional forms derived from physics and chemistry can be identified in the degradation of complex systems. This netSEM approach involves developing metrics, metrology, and tools to quantify, compare, and cross-correlate the response of PV modules, components, and materials to a variety of, and combinations of, environmental stressors (e.g., light, heat, and humidity) in both accelerated and real-world studies. With the incorporation of large-scale engineering epidemiological studies and statistical data analytics,

predictive and inferential/diagnostic mesoscopic evolution models which describe functional dependencies and degradation characteristics can be developed, using the netSEM package [64]. In this way, different regimes in degradation processes (i.e., initial stress, transitions due to induction period and change points, growth, and end-of-life failure) can be determined and modeled from the active degradation mechanisms. Then these can be joined together for a network of degradation mechanisms along a number of pathways. These network models highlight the key mechanisms and pathways that impact the performance of a PV system. This approach can ideally be helpful to provide valuable information for initial material selection and system design and performance prediction over its useful lifetime. In addition, the network and pathways of mechanisms provide to the materials designer, multiple opportunity areas for stopping or decreasing specific degradation pathways by neutralizing specific mechanisms, and their progression. The resulting enhanced performance, reliability, profitability can lead ultimately to successful and sustainable commercialization of products and hence widespread adoption of PV systems. It is to be noted that the following sections will include discussions for crystalline silicon (c-Si) PV modules unless otherwise noted.

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

3.2 Polymer-Related Failures in PV Modules A PV module consists of several components, including a front glass, solar cells with metallization elements such as silver grid lines, bus bars, string interconnects, embedded between two layers of encapsulant polymers, a polymeric backsheet, a junction box, and a frame (Fig. 3.2). The packaging components, such as the encapsulant and backsheet polymers, of PV modules deliver electrical insulation and environmental isolation and mechanical stability to the solar cells and interconnect metallization elements. Polymers used for packaging should therefore be elastomeric and flexible to protect the solar cells from mechanical impact and cyclic fatigue; stability against UV light, heat, and humidity, and high light transmission, particularly for encapsulant polymers, for cells to produce maximum power possible under illumination, and high dielectric strength, particularly for backsheets, for safe operation [66e72]. Ethylenevinyl acetate (EVA) is the most common encapsulant material used in crystalline silicon PV modules, yet silicone elastomer-, ionomer-, and polyolefin-based polymers are employed as alternative encapsulants [73e79]. A common backsheet structure consists of multilayer films such as a fluoropolymer layer, either

49

polyvinyl fluoride or polyvinylidene fluoride, on the outer (air-side) surface of the PV module (chosen for their high environmental stability [80]), a polyethylene-terephthalate (PET) core layer for high dielectric breakdown strength, and an EVA layer on the inner surface for good adhesion to the polymer encapsulant (i.e., FPE backsheet). In some backsheet applications, the inner EVA layer is replaced with a fluoropolymer layer (i.e., FPF-type backsheet), or the outer layer or both outer and inner layers are composed of stabilized PET films (i.e., PPE- or PPPtype backsheets), or other polymeric materials are used such as single-layer PET, polyamide, or polyimide films to avoid high cost of fluoropolymers [81e87]. Water vapor transmission rate (WVTR) and oxygen transmission rate (OTR) (or acetic acid transmission rate (AATR)) are also important properties for these packaging materials [88e99]. Using materials with very low WVTR, OTR, or AATR (i.e., unbreathable packaging construction) can help prevent penetration of moisture and oxygen into the module. It will also keep moisture or other gaseous degradation by-products, such as CO2 formed during lamination and acetic acid generated by hydrolysis of EVA polymer, trapped inside the module, and could cause further issues like delamination, blisters, or bubble formation [100e103].

Figure 3.2 Photovoltaic module components with metallization elements of the solar cells.

50

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Any failure to packaging polymers, such as delamination and cracking, either of the encapsulant or backsheet, not only exposes the modules’ active electrical components, such as the cells and their metallization, to environment but also compromises the electrical insulation and can cause a safety hazard, such as from electrical leakage. Therefore a robust lamination process along with appropriate stabilization of these polymers, particularly to UV light, heat, and humidity, is of critical importance. The right combination of the encapsulant and backsheet polymers is one of the key parameters to prevent failure and ensure safe and long-term operation of PV modules [104,105]. Although module manufacturing, specifically the lamination process, has an enormous effect on polymer properties and stability, such as the proper curing of EVA encapsulant and good adhesion of different layers in the module, and hence module reliability, here only the weathering-induced degradation will be discussed. An example of possible PV module performance degradation pathways, as tabulated comprehensively from the scientific literature, can be seen in Fig. 3.3. Polymer-related failures can be summarized in three main categories as follows.

3.2.1 Delamination and Mechanical Failures The interfacial adhesion failure at interfaces between the front glass, encapsulant, solar cell, and backsheet, called “delamination,” occurs when the adhesion between these layers has deteriorated. This

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

could be due to poor lamination process parameters, such as preheating or curing temperature and heating or cooling rate, vacuum pressure, the presence of contamination or absence of adhesion promoting additives, and weather degradation of polymers and adhesive layers, and residual thermal and mechanical stresses stored during lamination or over time during deployment [106e120]. Delamination, either from the front glass or backsheet side, or of the encapsulant itself, can cause increased oxygen and moisture ingress into the module followed by concomitant oxidation and corrosion of metallization elements [121e124]. When EVA, particularly with higher vinyl acetate content due to its increased solubility, undergoes oxidative and hydrolytic degradation, it releases free acetic acid that is known to react with silver-based grid lines and copper/tin bus bars and lead-based solder bonds and cell interconnects [125e129]. If the backsheet used is impermeable to acetic acid migration to the outside environment, then accumulation of acetic acid can increase the acidity of the microenvironment within the module and accelerate the metallization corrosion process [130]. In both cases, increased resistance due to corrosion of active electrical components can lead to significant power loss. Nonuniform optical coupling due to delamination at the glasseEVA or EVAecell interfaces can also result in reduced light transmission due to increased light reflectance and degradation of the antireflective coating on the cells and contribute to power loss [131]. Moreover, delamination, along with the associated increased moisture ingress and concomitant metallization

Figure 3.3 An example of photovoltaic module performance degradation pathways under heat, humidity and irradiance conditions. Individual and sequential impacts of stressors and the resulting interactions between degradation mechanisms can be seen. EVA, ethylene-vinyl acetate; PET, polyethylene-terephthalate.

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

corrosion is associated with current mismatch phenomenon giving rise to further power loss and to localized overheating (hot spot) issues due to the increased resistance [132,133]. This localized overheating can also hasten delamination, moisture ingress and hence corrosion, and discoloration processes as well as other failure modes such as burn marks. Delamination within the layers of the backsheet laminate, due to the degradation or loss of the adhesive interlayers, is one of the failure mechanisms seen in PV backsheet constructions [134e139]. Embrittlement, cracking, and crazing can also occur as a result of hydrolytic, thermal, and photolytic degradation of these materials as well as mechanical stresses caused by wind, hail, heavy snow load, and thermal cycling and can lead to deterioration of the mechanical and electrical insulation properties [140e148]. Crazing is the formation of very fine cracks on the surface layer induced by localized plastic deformation and can be seen as a precursor to embrittlement and crack formation upon sufficient growth over time. Chalking can also be observed in white-pigmented backsheet polymers as a sign of degradation. White pigments such as TiO2 are added to polymer to increase light scattering and the optical path length of photons in the polymer and to hinder photolytic reactions, thereby increasing the backsheet stability; however, extensive exposure to UV light and humidity can cause degradation of the exposed surface layer of the backsheet. This process can lead the polymer removal due to water and result in increased surface roughness. These failures are microscopic features usually characterized with surface reflectivity measurements [149e152].

3.2.2 Discoloration Discoloration of encapsulant and backsheet polymers occurs as a result of weather-induced thermooxidative and photo-oxidative reactions during outdoor deployment due to the formation of degradation by-products (i.e., chromophores) [153e159]. These chromophores cause strong broadband optical absorption in the UV and visible spectral regions, leading to a yellow (or even brown discoloration in case of EVA) appearance. Discoloration itself may not be directly related to power performance of PV modules; however, discolored EVA encapsulant is reported to cause a loss of light transmission to the

51

solar cells, leading to loss in power production by up to 0.5% a year by average [28] or up to 1% a year in subtropical climates [160], and this loss can be as high as 10% a year in case of mirror concentrated PV systems in desert conditions [161]. Upon prolonged exposure to environmental stressors, particularly to UV light and heat [162e170], intense discoloration can indicate deterioration of mechanical integrity of the polymer which might lead to delamination, embrittlement, and thus mechanical failure. In the case of embrittlement and mechanical failure, either of the encapsulant and backsheet, as mentioned in Section 3.2.1, power loss due to oxygen and moisture ingress into modules and corrosion of metallization and safety hazards due to compromised electrical insulation can be observed. Burn marks (or hot spot formation) can be seen on both encapsulant and backsheet due to overheating along bus bars, cell interconnect ribbons, and string interconnects observed in fielded modules [171e173]. Overheating can occur due to increased weather temperature along with increased resistance associated with metallization failures such as disconnected ribbons and solder bonds. Partial shading and the presence of nonfunctional cells can also lead to the formation of burn marks by inducing reverse or blocked current in part of a module, which could result in electric arc and fire [174e176]. Similar to burn marks, localized browning of EVA can be observed when EVA is in direct contact with corroded copperbased cell interconnects by acetic acid. Although burn marks are not directly related to degradation of the packaging polymers themselves, their presence can help identify failures within the modules’ active parts that may require serious attention. Another form of discoloration observed in modules is the formation of snail trails often found at the edge of solar cells due to degradation of silver-based metallization [177e179]. One pathway suggests that silver acetate, which forms by the reaction between silver/fritted glass paste and acetic acid generated by the degradation of EVA and oxygen is responsible for this color change [180,181]. The encapsulant or backsheet formulations can therefore contribute to or help prevent snail trail formation [182,183]. Although snail trail formation is known to have no direct effect on the power yield, they usually form along the cell cracks which cannot otherwise be detected visually [184,185].

52

D URABILITY AND R ELIABILITY

OF

P OLYMERS

3.2.3 Potential-Induced Degradation Potential-induced degradation (PID) is related to the mobility of ions, within the PV module, specifically between the cells and the modules’ front glass, caused by the modules’ voltage potential and leakage current, and can dramatically reduce the power production [186e191]. In this process, negative ions move away from the cells and positive ions move toward the cells from the glass and other packaging materials as seen in Fig. 3.4. The properties of packaging materials (e.g., the type of front glass, encapsulant, and backsheet) used, and environmental factors (e.g., heat and humidity), play important roles in PID formation [192e208]. PID is also examined at system level due to the effects of module and grounding configurations and inverter type on the module voltage potential and its sign, and at cell level due to the presence of the Si3N4 antireflection coating on the front side of the silicon PV cell; however, only the contributions from materials and environmental factors are considered here. Due to their high mobility, sodium ions (Naþ) have a strong influence on PID formation in modules with sodaelimeesilica front glass. Also, glass corrosion might occur as a consequence of accelerated dissolution of ions at both glasseEVA and EVAecell interfaces due to acetic acid generated by moisture-induced hydrolysis of EVA [209]. In this case, a backsheet with high acetic acid permeability (i.e., high AATR) can minimize this effect [210,211]. An encapsulant material with high moisture

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

impermeability (i.e., low WVTR) can be helpful as moisture ingress not only increases the possibility of hydrolytic degradation of EVA and metallization corrosion but also increases the conductivity which facilitates ion mobility and thus PID [212e214]. Therefore, an encapsulant material with high volume resistivity (i.e., impermeable to charged ionic carriers) can hinder the ion mass transfer and diminish the leakage current and PID [215]. However, not only resistivity of the encapsulant is affected adversely at elevated temperatures but the permeation rates of moisture and acetic acid are also temperature dependent [216]. One way to overcome this problem is the use of a thin impermeable layer, such as ionomer, polyethylene, or silicon dioxide (or silicon nitride) as a barrier layer between the glasseEVA and EVAecells interfaces to hinder the diffusion of ions [217e219]. Alternative materials that have high volume resistivity (i.e., lower polarity), low moisture permeability, and that obviate the production of acetic acid, when compared to EVA, can be used, examples being silicone elastomer-, ionomer-, and polyolefin-based encapsulants [220e222].

3.2.4 Discussion There are other failure modes (nonpolymerrelated) in PV modules such as frame and front glass breakage [223,224] and cracked cell, cell metallization, or disconnected cell and string interconnect ribbons [225e231] due to mechanical impacts, junction box and bypass diode failures [232,233], light-induced degradation [234e236], and

Figure 3.4 Schematic representation of potential-induced degradation. Ion motion within the module is shown as negative ions moving away from the cells and positive ions moving toward the cells from the packaging.

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

edge-seal degradation [237], but these are not a focus of this chapter. Considering the potential causes for power loss and safety problems, individual mechanisms cannot be studied in isolation because many degradation mechanisms can be active in a PV module over its lifetime.. The most significant degradation mechanisms observed in the fielded PV modules can be seen in Fig. 3.5. For instance, similar to encapsulation failure, frame or front glass breakage provides a rapid path for oxygen and moisture ingress, leading to delamination and degradation of polymeric materials, and subsequent formation of cell damage and metallization corrosion, and finally, power loss. All these components, their degradation, and the effects on the reliability of PV modules will be discussed in separate chapters within this book. In summary, failure of PV module components is not only detrimental for PV module performance, but it can also cause safety hazards such as electric shock or fire. In order to improve the reliability and service lifetime, understanding the origins of current failure mechanisms in modules is of essential importance. Various characterization methods and predictive models have been developed to detect and identify

53

failure mechanisms in degraded PV modules and materials during manufacturing, accelerated exposures, and outdoor service [238e273]. As depicted in this section, most failure mechanisms in modules are not because of a single exposure stressor or degradation of a single material or component, but the combination of multiple stressors and interactions between materials and mechanisms play critical roles. The following sections will provide an in-depth discussion of novel aspects in PV degradation and reliability research.

3.3 The Drawbacks of Standardized Testing The standardized qualification and certification tests used in PV industry allow for identifying manufacturing defects for the purposes of supply chain qualification and quality control; however, these standards are not designed to focus on mechanisms that cause degradation or performance loss in PV modules over its lifetime. These tests usually consist of single and/or constant stress variables applied at increased stressor intensity levels for a

Figure 3.5 Degradation mechanisms observed in the fielded photovoltaic (PV) modules for all years (dark colored) and the last 10 years (light colored). The bars are color coded by severity (red, high; green, medium; and yellow, low). Source: Jordan DC, Silverman TJ, Wohlgemuth JH, Kurtz SR, VanSant KT, Photovoltaic failure and degradation modes. Progress in Photovoltaics: Research and Applications 2017; 25(4):318e326. URL: http://onlinelibrary.wiley.com/doi/10.1002/ pip.2866/abstract, doi:10.1002/pip.2866.

54

D URABILITY AND R ELIABILITY

OF

P OLYMERS

relatively short amount of time (compared to the 25 year power warranty of a PV module) and are useful to determine infant mortality that occurs due to design, processing and manufacturing flaws; they are not capable of predicting long-term performance of PV modules, considering the complexity of realworld exposure conditions and the PV module when considered as a complex system itself. The appealing aspect of standardized testing is the shorter test duration and thus low cost of these “accelerated” tests, for a quick observation of products’ durability; yet PV module reliability and their actual degradation requires long time periods of exposure in the field. Unfortunately, service life prediction of many industrial applications relies upon accelerated testing to assure future reliability because of the very high cost and lengthy timescale of real-world testing as these can delay market introduction significantly. Accelerated testing only determines degradation under specific stress conditions, and it is then affiliated with the degradation that occurs in actual use conditions. Although standardized accelerated tests might seem very advantageous, such an approach carries large risks because of the intensified stressor levels due to the possibility of activating unrealistic degradation mechanisms not seen in real world instead of activating and accelerating realistic degradation mechanisms to a greater extent. Without a fundamental understanding of system properties, the stresses imposed by accelerated exposure conditions, or the degradation mechanisms, and pathways involved along with their dependence on stressors, accelerated testing can be misleading, leading to either neglecting important degradation modes, or adding cost to the module, to disable an unrealistic degradation pathway that will never be active under real-world conditions. Service lifetime performance prediction, as covered in Chapter 13 of this book, requires information on performance and degradation of PV modules and materials, which can only be obtained during deployment in the field over a long period of time. Degradation of a system or a material would depend on the number and degree of applied stressors (e.g., stressor levels can be constant, or varying with time, or even cyclic). So there could be more than one degradation mechanism contributing to overall performance. Due to the pass/fail nature of standardized tests, they typically do not meet the experimental requirements to produce sufficient data for analysis, nor do they approach the complexity of real-world

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

exposure conditions, and thus do not allow for determination of specific degradation mechanisms, rates, and pathways [274]. A survey of observed degradation mechanisms and a vast amount of experimental data obtained through both lab-based accelerated and real-world weathering exposures and their multivariate statistical analytics and crosscorrelation are therefore needed to provide a valuable, quantitative assessment of module reliability [275e280]. Understanding potential degradation mechanisms, their origins, and the complex interactions among them will help the PV community optimize study protocols for lab-based exposures so they can be more accurate and produce datasets that support the development of realistic predictive models of PV lifetime performance. Damp heat exposure is one of the qualification tests required by the IEC 621215 standard [281] in order to determine the performance of PV modules and module materials under excessive heat and humidity conditions. According to the test protocol, PV modules are aged under constant stressors of 85 C+ and 85% relative humidity (RH) for 1000 h and are required to survive with no more than 5% power degradation, no major visual defects, and no changes to insulation resistance and wet leakage current. This test might demonstrate premature failures that could be observed in the field upon installation, but there has been a strong tendency to correlate this test to service lifetime of PV modules. Forecasting of a lifetime of 25 years from a lab-based accelerated exposure of 1000 h (i.e., acceleration factor of w220) seems very practical; however, the damp heat exposure conditions applied to the PV module is much more aggressive than what would be seen in the real world. Under these exceptionally harsh conditions, the types of activated degradation mechanisms and their acceleration may strongly alter what the modules would experience over long-term field use. This test could result in manufacturers taking unnecessary measures to improve manufacturing and design parameters so as to pass this excessive standard’s requirements without actually improving the module’s real-world performance. For these reasons, the usefulness of this standard to predict service lifetime is a source of vigorous discussion in the PV community. One example [1] presents the misuse of this pass/fail “one-point data” test in a study of degradation of PET and polycarbonate (PC) polymers under 85 C and 85% RH. Using the Arrhenius relation, it was shown that even

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

though PET degraded more rapidly than PC under the applied stressors, when extrapolated to the ambient end-use temperature (25 C), PET survived better than PC because they had different activation energies and both lasted an order magnitude longer lifetime than required by the standard. Since the activation energies are also temperature dependent, the use of the activation energy determined at 85 C for a lifetime at ambient temperature raises serious questions. Another study [282] showed that damp heatedriven moisture concentration in the encapsulant between the cell and glass layer is two times higher than that the real-world weathering can accumulate over 20 years using measured microclimate data and FEM (finite element analysis)based simulations. For the hydrolysis of PET polymer [283], 1000 h damp heat exposure was found to cause degradation that would occur over 150 years of real-world exposure in Bangkok, Thailand, one of the hottest and most humid locations in the world. Similarly, in Ref. [284], 1000 h of damp heat was found to correspond to 16 years in a tropical climate zone or more than 100 years in a middle European climate zone for the degradation of polyurethane-type encapsulant material in a rear insulated PV modules, in which temperature can reach 60 to 80 C. For free-standing modules, no correlation was found between damp heat and realworld exposure. In Ref. [285], damp heat exposure times were shown to differ from w500 to w2000 h to achieve a given time to failure for PV module performance over 25 years of field exposure at different locations in the world. The time to failure due to metallization corrosion was demonstrated in Ref. [286] under different damp heat conditions and compared to those in different locations. Because of its higher activation energy and relative humid conditions, the highest failure rate obtained under damp heat conditions suggested a lower overall risk of a field failure for hot and humid locations. While physical acceleration models developed in the study were consistent for mild and hot and humid locations, significant deviation was observed for dry locations. It was noted in Refs [287e289] that failures driven by real-world exposure conditions are not replicated by the current damp heat standard. So new test protocols that involves UV light, mechanical loading, and temperature cycling are proposed to adequately simulate observed field failures. In summary, uncertainty arising from results of the damp heat test is a serious concern, most of all when one considers the

55

diversity of different locations, climatic zones, and mounting configurations that PV modules are exposed to in the real world.

3.4 The Lifetime and Degradation Science Approach PV modules suffer loss of performance during operation, and their service lifetime depends on the type and rate of degradation mechanisms activated under the real-world conditions they are exposed to. Two of the main targets of PV reliability research are to develop appropriate and robust models for the purpose of accurately predicting the reliability and lifetime performance of PV modules, and to develop lab-based accelerated exposures that closely simulate real-world conditions and their impact on PV modules. For a detailed understanding of PV system degradation, scientific investigations that span a wide range of physical and chemical PV degradation mechanisms are critical. However, in traditional PV reliability studies, observed degradation mechanisms usually are considered in isolation from a complex systems perspective. Data science and statistical analysis applied to study protocols (tests) encompassing exposure conditions, evaluations (measurements using various tools), and data-driven modeling can therefore be used to find connections between important degradation mechanisms suggested by domain knowledge (i.e., based in physics and chemistry) and discover their individual and cumulative effects on system performance. This methodology would help reveal the complexity of many factors contributing to degradation of PV performance in a systematic way and ultimately reach reliable predictions about service lifetime. The L&DS approach aims to obtain quantitative information about degradation mechanisms and rates under applied stressor (exposure) conditions as depicted in Fig. 3.6. In this approach, relationships between applied stressors, observed mechanisms, and cumulative responses are determined quantitatively using statistical data analytics, and the relationships among these are represented using a network-based degradation pathway model. In the field, exposure stressors and stressor levels are uncontrolled, unpredictable, and variable over time. One of the few ways that real-world exposure conditions are holistically categorized is by the Ko¨ppeneGeiger climatic zones system [8] that has

56

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 3.6 Elements of lifetime and degradation science for photovoltaics.

been developed since the 1884 work of Ko¨ppen [290,291] Due to various interactions caused by all types of stressors, there will be competing mechanisms and hence complex degradation pathways in the system under investigation. Furthermore, stressors such as light cycles during day and night, temperature, and RH changes during the day or throughout the year, and mechanical stresses due to rain, hail, snow load, and wind, can induce physical, chemical, and mechanical degradation which cannot be anticipated in accelerated testing. Single exposure stressors usually lead to simplistic assumptions since stressors often possess synergistic effects which can accelerate (or decelerate) degradation rates and alter degradation mechanisms and hence pathways. A multifactor test with a wide variety of stress levels and types of stressors, preferably with cyclic conditions, can provide a better understanding of these synergistic effects. The comparison of response variables under a single stressor and multiple stressor exposures can also clarify whether the mechanistic synergies arise only for combined (simultaneous) stressors or also accumulate under sequential exposures to those stressors. Overall degradation or performance loss will then depend on the integrated and mutual effects of all stressors applied to the system. Running well-designed lab-based accelerated study protocols and real-world study protocols in parallel and using quantitative cross-correlation methods is therefore essential for L&DS of PV systems. Because of simplicity, low cost, and short timescales, accelerated testing is usually preferred over real-world testing; however, long-term effects of real-world conditions on PV systems cannot easily be

replicated in laboratory conditions. The stressor conditions in accelerated testing are too harsh, leading to overengineering the system since these mechanisms are not even active in real-world conditions. And if the stressor conditions are too simplistic in the accelerated test, then the performance and durability of the system will be overestimated, misleading the community on its capabilities. So a better scientific and quantitative understanding of PV degradation mechanisms is key to designing lab-based exposures that simulate longterm behavior of PV materials and systems. This would allow the PV community to identify and incorporate necessary improvements in industrial processes in shorter time frames and help advance the reliability of PV systems. In the stress and response () framework of L&DS, timewise observational data are collected on the degradation of materials or systems over multiple and cyclic stress conditions. Relationships between many different stressors, mechanisms, and responses and their quantitative cross-correlation are then established, instead of the traditional approach of a simplified “acceleration factor” that applies to a specific test condition. So all of the available degradation mechanisms and rates can be integrated to provide a comprehensive, scientific, and quantitative understanding of the performance of the system over lifetime. And for the specific system and exposure conditions used (the study protocol), this information can be represented as a network diagram (netSEM) of the relationships between the stressors, the mechanisms, and responses. This is refereed to as the degradation pathway diagram for that system under

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

those specific exposure conditions. The scientific understanding of the lifetime performance of a material or system can be gleaned by comparing results of real-world exposures to the stress and response framework. If good cross-correlation is observed between the real-world and accelerated studies and the system’s responses, then those lab-based accelerated test, summarized in the study protocol, can be beneficial when predicting the degradation behavior of a system in actual use environments, instead of requiring the lengthy, high-cost real-world exposures. For robust adoption of feasible PV power systems as a competitive green energy resource, reliability is of great concern. Lifetime and degradation studies will enlighten existing physical and chemical mechanisms, and open paths forward to further research.

3.4.1 Statistical Data Analytics for Lifetime and Degradation Science 3.4.1.1 netSEM Modeling Approach In order to assess the results of L&DS studies and thus establish degradation models and pathways, a

57

longitudinal stepwise experimental design with a large body of data and its analysis through statistical methods are of great importance. SEM [62,63,292] is a common technique used in social sciences to derive and map casual relationships between latent (unobserved) variables from observable variables through mathematical systems of equations as shown in Fig. 3.7. However, traditional SEM uses linear models to evaluate these relationships and lacks quadratic, exponential, and logarithmic models that are naturally observed in physical and chemical processes. For this reason, the traditional SEM methodology has been adapted to introduce nonlinear models and exploratory data analysis in order to determine variables related to particular degradation mechanisms. This new methodology, referred to as netSEM, is now available as an open source code package in the R programming language [47,53,64]. netSEM is semisupervised, whereby domain knowledge is used to supervise the stepwise variable selection and model development, and generalized so as to incorporate nonlinear models in the analysis.

Figure 3.7 netSEM pathway diagram and systems of equations for the lifetime and degradation science approach. S1 and S2 are the stress variables, M1 and M2 are the mechanistic tracking variables and latent variables, where tracking variable is the direct measurement used to quantify latent variable. R1 is the performance level response variable. The coupling strength (bi, j) between two variables is the coefficient vector of the fitting model that predicts variable i from j. Contributing factors to overall degradation can be determined using the framework.

58

D URABILITY AND R ELIABILITY

OF

P OLYMERS

With the help of netSEM analysis, predictive () and mechanistic/inferential () data-driven models can be determined for the temporal evolution of degradation in the system under study. In this approach, the degradation of a systemlevel performance response (R), typically a variable of most interest for the system under study, will be a result of numerous contributing factors. netSEM models can be built using two different statistical principles to aid in variable and model selection [293]: models fitted using Principle 1 are fitted in a Markovian spirit, while models fitted using Principle 2 are fitted in a multiple regression manner [47]. Principle 1 determines the univariate relationships in the spirit of the Markovian process [294,295]. This means the relationship between each pair of system variables is determined with the Markovian property that assumes the value of the current predictor is sufficient in relating to the next level variable (i.e., the relationship is independent of the specific value of the preceding-level variable to the current predictor) given the current value. It is important to note that the Markovian nature of Principle 1 netSEM modeling of these univariate relationships ignores the simultaneous impact of other variables in the system and treats the relationship as occurring solely between the two variables in question. We therefore also developed Principle 2 which resembles the multiple regression principle in the way multiple predictors are considered simultaneously. Specifically, for Principle 2 the first-level predictors to the system level variable, such as, time and unit level variables, acted on the system level variable collectively by an additive model Pmax ¼ fi(time) þ fk(IR2) þ . þε where fi’s are appropriate (parametric) function fits. The analysis of a system involves six essential steps: 1. Designing a study protocol with a comprehensive set of exposures, evaluations (measurements that produce the data to be analyzed) and an appropriate number of time steps (typically six to nine time steps) at which evaluations are performed and data acquired. Multiple coincident observations of variables related to responses of the system under study to describe the degradation mechanisms are essential over the entire course of the study for accurate statistical analysis.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

2. Selection of variables by domain knowledge. These variables are a stressor(s) (S), mechanistic tracking and latent variables (M), and a system level performance response (R). 3. Model fitting of every variable using a pool of linear and nonlinear functional forms. The analysis starts from the system level performance response (R) variable and treats every other variable as a predictor, and proceeds using stepwise regression [296]. 4. Determination of the statistical relationships between variables. Statistical diagnostics are performed on each of the functional forms between two variables, to determine the best functional form using the calculated adjusted R2. 5. Sequential analysis of step 3 and step 4 until all the relationships are found and the best functional fits are achieved. If the adjusted R2 value of a relationship is above some set threshold value, then the relationship can be considered to be strong. 6. Analysis of the obtained degradation pathways using domain knowledge. If the resulting relationships suggest any degradation mechanisms known from domain knowledge, then they might imply potential causal factors and pathways for the degradation. Each relationship is defined by an equation, and the resulting degradation pathway will then be described by a system of equations for the system as shown in Fig. 3.7. The advantage of netSEM analysis is that statistically significant relationships connecting applied stressors to degradation mechanisms and final responses of interest can be evaluated and mapped into a degradation pathway of the system under investigation. This new netSEM methodology has been coded using R language [297] package for automated, fast, and easy analyses of big datasets [298]. The analysis can be run throughout the study at any point in time and rerun when there is additional data from further exposure and evaluation steps in the study. The evolution of degradation with additional data along the ongoing study can highlight the different regimes of the degradation process and support the existing pathway with increased predictive power or bring out new aspects for the development of new pathways. The application of netSEM

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

methodology to long-term data of fielded PV systems could also facilitate the design of lab-based accelerated test methods by comparing the degradation experienced under the two conditions.

3.4.1.2 Fixed, Random, and Mixed Effects Regression Modeling Approach Linear regression models are often preferred to quantify relationships between variables of interest. However, in a longitudinal experimental design with multiple coincident observations of variables at multiple points in time, and specifically with multiple number of predictors, model selection is a challenge. Models, in this case, are expected to consider all the variables in a multilevel way and the interactions between them [299]. Considering the material, component, and system-level degradation in PV modules under multifactor stressor conditions, multivariate linear regression modeling is a wellsuited method for the L&DS of PV systems. Using fixed-effects [300] and mixed-effects modeling [301], in which fixed and random effects are included, can allow for multilevel modeling of variables and determination of contributing factors to degradation [302]. The terminology of fixed effects, random effects, and mixed effects, is not well standardized, so one should carefully clarify the different approaches [303]. Service lifetime prediction of PV modules, performed using results of lab-based standardized tests, is not technically feasible because of the presence of multiple and variable stress conditions in the real world. Owing to their ability to model multilevel structured and longitudinal experimental design, multivariate regression modeling approach, however, can be extended to model realworld scenarios. In multilevel models, covariants are structured in a hierarchical way so that there is more than one level in the study design. For instance, repeated measurements of multiple samples of different materials (lower level) used in different brands of PV modules (upper level) that are stressed under different conditions can be studied and modeled at the same time. In this case, all the individual samples will have their own variation, and depending on the between-sample variation, two different approaches can be utilized.

59

Fixed-effects modeling is implemented for multivariate analysis of repeated measurements of multiple samples when the variance between samples is small. However, in case of larger variance between samples, mixed-effects modeling that include both fixed effects and random effects can be advantageous. Random effects are described as sample-specific effects related to unobserved or unmeasured factors in the study (i.e., larger variance due to measurement uncertainty or different characteristics of individual samples). Incorporation of random effects improves the overall explained variance and hence increases the statistical significance of the models. The multiple regression model then takes the general form as shown in Eq. (3.1). y ¼ b0 þ b1 x1 þ b2 x2 þ / þ bn xn þ b1 z1 þ b2 z2 þ / þ bn zn þ ε (3.1) where y is the response variable, b0 is the intercept of the regression, b1 through bn are the fixed effect coefficients (i.e., parameter estimates), x1 through xn are the fixed effect variables, b1 through bn are the random effect coefficients, z1 through zn are the random effect variables, and ε is the error term.

3.4.1.3 Multivariate Multiple Regression Modeling Approach The linear multiple regression modeling approach can be expanded to include multiple response variables when several metrics are under study for the L&DS of PV systems. The multivariate multiple regression (MMR) approach assumes that all of the response variables can be described by the variation same set of covariants in the study. In the context of the hierarchical study mentioned in Section 4.1.2, this methods allows repeated measurements of multiple properties (multivariate) of multiple samples of different materials (lower level) used in different brands of PV modules (upper level) that are stressed under different conditions to be studied and modeled at the same time. In the MMR model, the response vector y from univariate linear regression is replaced by an n  m

60

D URABILITY AND R ELIABILITY

OF

P OLYMERS

matrix of responses Y. Generally, a MMR model can be represented in Eq. (3.2), Y ¼ XB þ E;

(3.2)

where n is the number of observations, m is the number of response variables, p is the number of predictor variables, Y is an n  m matrix of responses, B is a (Pþ1)  m matrix of regression coefficients, and E is an n  m matrix of error terms [60].

3.4.1.4 Parallel Factor Analysis Modeling Approach Chemometric techniques are commonly used to transform data in higher order space (>1) (i.e., spectroscopic data, to a space of fewer dimensions). The chemometric techniques that have been proven useful for dimensional reduction of spectral data include principal component analysis (PCA) and parallel factor analysis (PARAFAC) for twodimensional (e.g., spectral or xey data) and data of greater than two dimensions (e.g., hyperspectral, xeyez data, etc.), respectively. These techniques can be used within and frameworks to define mechanistic and response variables when the data are not initially low dimensional. PCA and PARAFAC perform linear mapping of the data to the one-dimensional (point data) space such that the variance of the dimensionally reduced data is maximized, and the information retention is maximized after the transformation [304e306]. Historically, PARAFAC has most commonly been applied to excitationeemission matrix (EEM) fluorescence spectra [307]. PARAFAC is a mathematical procedure that decomposes an M dimensional array into the summation of the outer product of M vectors. PARAFAC applied to EEM spectra (EEMPARAFAC) provides an estimate of the number of fluorophores as well as estimates of the excitation and emission spectrum of those fluorophores. It also provides the relative concentration of each fluorophore in each sample. A general PARAFAC model(s) can be written as shown in Eq. (3.3), where xijk is the fluorescence intensity of the kth sample at the ith emission wavelength and the jth excitation wavelength, ain represents the estimated emission spectrum of the fluorophore(s), bjn represents the estimated excitation spectrum of the fluorophore(s), ckn is the relative of the

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

fluorophore(s) in sample k, and eijk represents the error terms or residuals [61]. xijk ¼

N X

ain bjn ckn þ eijk

n¼1

(3.3)

i ¼ 1; .; I; j ¼ 1; .; J; k ¼ 1; .; K

3.4.2 L&DS Case Studies 3.4.2.1 PV Degradation Pathway Model under Damp Heat Exposure In a particular study conducted by researchers at Underwriter’s Laboratory (UL) [308], modules were exposed to damp heat conditions for 4000 h and the degradation in PV power output was measured at specific time intervals, or time steps. In addition, some modules were withdrawn from the study at these time steps to produce a retained sample library with samples available for each time step. These retained samples were dismantled destructively in order to determine the changes occurred in the module materials. With exposure time as a proxy for damp heat exposure stress (S) variable and Pmax (maximum power generated by the modules) being the performance level response (R) variable, several mechanistic variables (M) obtained through chemical evaluation methods performed on the subsystem components were used in the netSEM analysis [47]. The analysis revealed main degradation pathways as shown in Fig. 3.8. The strong relationships between the applied stressor, the infrared spectra (IR) signals of the EVA encapsulant, and the amount of free acetic acid measured through thermogravimetric analysis of EVA, indicate degradation was occurring through hydrolysis of EVA polymer under damp heat conditions. The effect of these two mechanistic variables on the loss in Pmax suggested a path of EVA hydrolysisedriven power loss in PV modules. This particular path was attributed to the metallization corrosion of silver grid lines and bus bars screen printed on solar cells caused by acetic acid content generated by the hydrolysis of EVA. The power loss exhibited an induction period (i.e., delayed onset of degradation) followed by a sudden drop in power after a certain point in time, i.e., a change point phenomenon [309,310], suggesting the presence of different degradation regimes due to damage accumulation over time. The results of this study later led to many further research directions,

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

61

Figure 3.8 PV degradation pathway model under damp heat exposure. Time is the main stressor as a proxy to damp heat exposure, Hac is the formation of acetic acid measured through thermogravimetric analysis (TGA), EVA_hyd stands for EVA hydrolysis, which is tracked by infrared signal corresponding to that, noted by IREVA, and Pmax is the maximum power generated by the PV modules. Models are as follows: SL is simple linear, SQuad is simple quadratic, Quad is quadratic, Exp is exponential, and Log is logarithmic, CP is change point, and nls is nonlinearizable exponential. Adjusted R 2 values are given for the power of functional fitting for each relationship.

as included in the following case studies, specifically for the degradation of polymeric materials used in PV modules under accelerated weathering exposures, extending the L&DS approach to subsystem components.

3.4.2.2 Mini-Module PV Degradation Pathway Model under Damp Heat Exposure The Carissa Plains power plant, the largest PV module installation of its time originally rated at 5.2 MW, was built around 1984. However, it showed over 40% power degradation by 1989 with an average rate of 7% per year, as opposed to expected loss of 1% per year, and was decommissioned soon after [311,312]. This unfortunate and drastic failure raised serious concerns about long-term reliability of PV systems. The main problem was initially believed to be related to EVA degradation due to the dark brown color of the PV modules (i.e., discoloration of EVA as a result of thermo-photo-oxidative degradation). Therefore, the degradation of EVA took special attention after this incident although later studies [313e315] on the recovered fielded modules discussed other possible causes of power degradation such as cell mismatching due to inappropriate use of concentrating mirrors, solder joints, wiring, and inverter problems.

To reveal the details of EVA hydrolysiseinduced PV power degradation, a new study, based on the netSEM analysis of this particular degradation path, was pursued [316]. This study used EVAencapsulated four cell-mini-module constructions in order to investigate the effect of acetic acid formation on the mentalization corrosion of the screen printed silver conductive lines and its relation to power performance under damp heat conditions. The acetic acid is known to react with the silver lines causing increased resistivity and hence decreased conductivity between the silver lines and the cell surface as explained in subsection 2. The hydrolysis of EVA was confirmed by laser confocal Raman microscopy in the damp heate exposed mini-module constructions without the need for disassembly. A broad fluorescent peak can be found around 1340 cm1, which is a signature of various degradation products of EVA hydrolysis. The impact of acetic acid formation on the electrical performance of each individual cell was determined via electroluminescence (EL) imaging and power performance of the modules. The power loss was ascribed to the loss of conductivity of the silver grid lines caused by the formation of corrosion byproducts such as silver acetate at the grid line and emitter interface reducing the conductive contact area [317,318]. Strong relationships between the chemical signals of EVA hydrolysis, quantitative EL imaging, and power performance data, obtained

62

D URABILITY AND R ELIABILITY

OF

P OLYMERS

through the netSEM analysis as shown in Fig. 3.9, confirmed this pathway with statistically coupled causal sequence of degradation mechanisms.

3.4.2.3 netSEM Modeling of PET Degradation PET polymer, with its high dielectric breakdown strength and low cost, is a critical material used in PV backsheets, acting as electrical and mechanical barrier. The degradation of PET-based backsheet was one of the pathways observed in the UL study, as presented in subsection 4.2.1, that led to loss of wet insulation resistance and hence power loss. In this case study, the durability of various stabilized PET polymer films were examined under multifactor light, heat, and humidity conditions [58,319]. These exposure are industrial standards [320] applied to test polymer durability. Exposure conditions are as follow: 1. ASTM G154-Cycle 4: cyclic exposure of 8 h of UVA light at 1.55 W/m2 at 340 nm at 70 C and 4 h of condensing humidity at 50 C in dark. 2. Modified-ASTM G154-Cycle 4: constant exposure of UVA light at 1.55 W/m2 at 340 nm at 70 C.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

For performance level response variables, yellowing and haze formation of these stabilized films were selected as a measure of physical and chemical changes under the applied stressors. Stepwise point in time data extracted from specific signals of UV-Vis and infrared spectroscopy were identified as mechanistic variables of the degradation. Strong coupling was found between these mechanistic variables and performance level response variables. Particularly, in the presence of only UV light as seen in Fig. 3.10, yellowing was significantly related to chain scission of the polymer backbone as measured through IR. The broadening of carbonyl band at 1675 cm1 was used as a sign of carboxyl end group generation. However, in the presence of cyclic heat and condensed humidity, in addition to UV light, strong haze formation was observed as depicted in Fig. 3.11. In both cases, aging-induced crystallization was observed through the IR absorptions of the trans oxyethylene glycol band at 975 cm1. So there existed one important degradation pathway: yellowing in the presence of UV light followed by haze formation in the presence of humidity. In this situation, haze formation that occurred as a bulk volumetric change, was assumed to be a precursor to loss of mechanical properties of the polymeric films. The relationship found between crystallization and haze formation also supported this idea. It is interesting to note that although yellowing

Figure 3.9 Mini-module PV degradation pathway model under damp heat exposure. Time is the main stressor as a proxy to damp heat exposure, EVA_FL is fluorescent degradation product of EVA, which is represented by a broad peak at 1340 cm1 in confocal Raman spectra. Hac is the acetic acid formation that will lead to metallization corrosion and is determined from the overall brightness of electroluminescence (EL) image, and Pmax is the maximum power generated by the PV modules. Models are as follows: SL is simple linear, SQuad is simple quadratic, Quad is quadratic, Exp is exponential, Log is logarithmic, CP is change point, and nls is nonlinearizable exponential. Adjusted R2 values are given for the power of functional fitting for each relationship.

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

63

Figure 3.10 Degradation pathway of UV-stabilized polyethylene-terephthalate under the modified ASTM G154-Cycle 4 exposure. Time is the main stressor as a proxy to modified ASTM G154-Cycle 4, AbsEdge is the fundamental absorption edge represented by absorption at 312 cm1 in UV-Vis spectra, UVStab is the bleaching of the UV stabilizer represented by absorption at 340 cm1 in UV-Vis spectra, Cryst is the infrared signal at 975 cm1 corresponding to change in crystallinity, ChainScs is the infrared signal at 1675 cm1 corresponding to chain scission, and YI is the yellowness index of the PET samples. Models are as follows: SL is simple linear, SQuad is simple quadratic, Quad is quadratic, Exp is exponential, Log is logarithmic, CP is change point, and nls is nonlinearizable exponential. Adjusted R 2 values are given for the power of functional fitting for each relationship.

was thought to be induced only by UV light, the presence of humidity along with UV light accelerated the formation of yellowing for the same amount of applied photodose. All these complex relationships were captured quantitatively by the netSEM analysis as seen in Fig. 3.11. Not only did a larger number of variables became active but also the correlations between the variables got stronger in the presence of multiple stressors and cyclic exposure conditions. This signifies the importance of synergistic effects of multiple stressors on degradation mechanisms, altering reaction rates and thus pathways. The role of the UV light stabilizer was notable in the UV light exposure, but not very effective. During the induction period, the rate of initial yellowing was slow; however, the rate accelerated with a strong linear dependency afterward. The stabilizer was degraded and consumed after a very short exposure time and left the polymer susceptible to damaging effects of UV light. Similarly, when compared to unstabilized polymers, the protection for hydrolytic attack was not found to be strong enough to withstand the intensified levels of stress conditions. The catalyst used during synthesis and the amount and type of

by-products formed in the initial polymers, such as carboxyl and hydroxyl end groups and diethylene glycol side groups, differ because of different synthesis routes in PET polymerization. These factors also have a strong influence on the stability of the PET films against UV light and humidity.

3.4.3 Fixed and Mixed Effects Regression Modeling of PET Degradation In addition to netSEM analysis, univariate regression analysis was performed in order for developing predictive models of temporal evolution of degradation of PET films under the applied multifactor exposure conditions. The study was suitable for this analysis due to its multilevel longitudinal experimental structure: multiple number of samples of different polymer grades tested and evaluated timewise under multiple number of exposures. Predictive models of yellowing and haze formation, as performance level response variables, were demonstrated using fixed- and mixed-effects modeling approaches [57].

64

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 3.11 Degradation pathway of UV-stabilized polyethylene-terephthalate under the ASTM G154Cycle 4 exposure. Time is the main stressor as a proxy to modified ASTM G154-Cycle 4, AbsEdge is the fundamental absorption edge represented by absorption at 312 cm1 in UV-Vis spectra, UVStab is the bleaching of the UV stabilizer represented by absorption at 340 cm1 in UV-Vis spectra, Cryst is the infrared signal at 975 cm1 corresponding to change in crystallinity, ChainScs is the infrared signal at 1675 cm1 corresponding to chain scission, Haze is the haziness, and YI is the yellowness index of the PET samples. Models are as follows: SL is simple linear, SQuad is simple quadratic, Quad is quadratic, Exp is exponential, Log is logarithmic, CP is change point, and nls is nonlinearizable exponential. Adjusted R 2 values are given for the power of functional fitting for each relationship.

Color change is a bulk property as the chromophores form homogeneously through the volume because of their very small molecular sizes, causing small variance between samples and repeated measurements. For this reason, the fixed-effects modeling approach was an appropriate choice for modeling of the yellowing of these films. On the contrary, haze formation is rather a random and localized process, and therefore the mixed-effects modeling approach was preferred. Partial crystallization due to hydrolytic degradation and microcrack formation due to thermal and mechanical stresses induced by cyclic

temperature are the main reasons for the observed haze formation. These locally distributed entities result in strong scattering of light causing measurement sensitivity and thus greater variance between samples and repeated measurements. The fitted models are shown in Fig. 3.12 for yellowing using fixed-effects and in Fig. 3.13 for haze formation using mixed-effects modeling approaches. It is seen that both models follow the experimental data quite well as their significance are evident from the calculated adjusted R2 values of 0.98 for yellowing and 0.90 for haze formation. The random

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

65

Figure 3.12 Predictive modeling of change in yellowness index via fixed-effects linear regression modeling. Unstab is unstabilized PET, HydStab is hydrolytically stabilized PET, and UVStab is UV-stabilized PET. PET, polyethylene-terephthalate.

effects, caused by each individual sample in the haze formation model, were found to improve the explained variance very well. Although adjusted R2 is a common metric for judging the predictive power of regression models, in case of a over-fitting problem, adjusted R2 may lead to inaccurate measure of model prediction. Instead, predictive R2 with its resampling method can yield more precise results. In this new approach, while a model is developed using the train data, its ability to predict a new observation is tested on the test data. In order to assess true measure of predictive quality of these models, predictive R2 values were determined using the leave-one-out cross-validation procedure. Even though the models were found to fit the experimental data very well according to adjusted R2 values, calculated predictive R2 values indicated minor over-fitting problems. So the use of predictive

R2 was explored successfully and shown to be valid technique as an accurate measure of predictive power.

3.4.4 Multivariate Multiple Regression Modeling of PET Degradation MMR was conducted in addition to netSEM and univariate regression to develop predictive models of the temporal evolution of several response variables for PET film degradation under the applied multifactor exposure conditions. The study utilized a similar multilevel longitudinal experimental structure to that described in Section 4.3: multiple samples of different grades of PET tested and evaluated timewise under multiple exposures. Predictive models were developed to account for discoloration,

66

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 3.13 Predictive modeling of change in haze formation via mixed-effects linear regression modeling. Unstab is unstabilized PET, HydStab is hydrolytically stabilized PET, and UVStab is UV-stabilized PET. PET, polyethylene-terephthalate.

gloss loss, and haze formation simultaneously under the various exposure conditions and configurations [60]. In addition to the previous descriptions of color change and haze formation (Section 4.3), gloss loss occurs as the surface of the PET roughens due to microcrack formation at the surface and localized embrittlement due to crystallization. Roughening at the surface increases reflective light scatter and a decrease in specular reflection, which is measured as gloss loss. The model for clear PET degradation under outdoor exposure conditions in Arizona and Florida is superimposed on the observed values in Figs. 3.14e3.16; outliers have been removed from the plots and subsequent analysis. The superimposed

plots support the reliability of the models as predicted values trace the experimental data reasonably well. The model significance is evident from the calculated adjusted R2 values of 0.93 for discoloration and 0.94 for gloss loss and 0.96 for haze formation. Results of model interpretation have provided insights into PET weathering. Photodose was found to be the prime source of degradation stress for PET. Samples degraded more quickly in Florida than Arizona, which implies that moisture and humidity play a role in increasing the degradation rate of PET. Direct moisture contact was found to increase the degradation rate of PET, which can be observed by comparing between the red and blue model curves from the model results (samples exposed open to the

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

67

Figure 3.14 The change in DE clear grades of polyethylene-terephthalate under outdoor exposures with the model curves superimposed on the data. Points represent the measured data and dashed lines represent the models.

elements vs. samples exposed under a glass cover sheet.)

3.4.5 Parallel Factor Analysis of Fluorescence Spectra for PET Degradation PARAFAC was applied to 63 EEM fluorescence spectra to quantify the evolution of degradation products (monohydroxy-terephthalate and dihydroxyterephthalate) under accelerated exposures [61]. Fluorescence spectroscopy is strongly sensitive to changes in chemical structure and can thus be used to

study the chemical changes due to PET weathering. A three-component PARAFAC was developed and interpreted to identify and evaluate the temporal evolution of PET and its two fluorescent degradation products. An example of the EEM fluorescence spectra, with first- and second-order Rayleigh scatter removed (diagonal lines), for unstabilized PET under three exposure conditions is shown in Fig. 3.17. The study included unstabilized PET, UV-stabilized PET (0.20% Tinuvin 360), and a slightly TiO2-loaded film (0.2% pigment volume concentration). The exposures included ASTM G154-Cycle 4 for 1 week time

Figure 3.15 The change in DGloss60 of clear grades of polyethylene-terephthalate under outdoor exposures with the model curves superimposed on the data. Points represent the measured data and solid and dashed lines represent the models.

68

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 3.16 The change in DHaze of clear grades of polyethylene-terephthalate under outdoor exposures with the model curves superimposed on the data. Points represent the measured data and solid and dashed lines represent the models.

exposure intervals (G154), a modified version of ASTM G154-Cycle 4 that omitted the condensing humidity stage for 1 week time exposure intervals (mG154), and standard damp heat exposure (85 C and 85% RH) for 2 week exposure intervals (DH). The three-component PARAFAC model yielded estimation of the excitation spectra, emission spectra,

and relative concentrations of the three fluorophores that contributed to the fluorescence signal in Fig. 3.17. The estimated excitation and emission spectra of the first two components are shown in Figs. 3.18 and 3.19. The estimates were compared with known fluorescence spectra to assign the components to relevant chemical structures: Component

Figure 3.17 The temporal evolution of the corrected excitationeemission matrix spectra of unstabilized polyethylene-terephthalate at three exposure steps. Row “a” corresponds to mG154 exposure, row “b” to G154 exposure, and row “c” to DH exposure.

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

Figure 3.18 Parallel factor analysis estimate of the excitation and emission spectra of Component 1. The solid orange line corresponds to the emission spectrum and the dashed purple line to the excitation spectrum.

69

Figure 3.20 Normalized relative concentration of Component 1 (monohydroxy-terephthalate units) over time under each exposure for the three grades of polyethylene-terephthalate (PET). Solid green lines correspond to ASTM G-154 Cycle 4 exposure, short-dashed orange lines to modified-ASTM G-154 Cycle 4 exposure, and long-dashed purple lines to damp heat exposure. Circles correspond to PET, triangles to PET-TiO2, and squares to PET-UVS.

Figure 3.19 Parallel factor analysis estimate of the excitation and emission spectra of Component 2. The solid orange line corresponds to the emission spectrum and the dashed purple line to the excitation spectrum.

1 was assigned to monohydroxy-terephthalate units and Component 2 was assigned to dihydroxyterephthalate units. Component 3 was assigned to PET; however, these data have been omitted to focus on the degradation products. Figs. 3.20 and 3.21 show the normalized concentrations of these degradation products that result from hydroxy substitution at the PET aromatic ring. UV light is necessary to drive this degradation mechanism, and the presence of moisture (G154) promotes further degradation. The figures also permit comparison between the different grades of PET and show that the slightly TiO2-loaded film showed the highest resistance to degradation and formed the lowest concentrations of the degradation products.

Figure 3.21 Normalized relative concentration of Component 2 (dihydroxy-terephthalate units) over time under each exposure for the three grades of polyethylene-terephthalate (PET). Solid green lines correspond to ASTM G-154 Cycle 4 exposure, short-dashed orange lines to modified-ASTM G-154 Cycle 4 exposure, and long-dashed purple lines to damp heat exposure. Circles correspond to PET, triangles to PET-TiO2, and squares to PET-UVS.

70

D URABILITY AND R ELIABILITY

OF

P OLYMERS

3.5 Conclusions The standard accelerated tests are the basis of PV module qualification for today’s market introduction of products, but they only provide a quick screening for potential design problems and detection for failure mechanisms in a specific environment. Since they focus on the effect of particular stressors and degradation mechanisms on module performance, it is not viable to say that a module that passes the standard tests with binary outcomes (i.e., pass/fail) will operate a long service lifetime in the field. However, successful lifetime prediction requires scientific understanding of all degradation mechanisms and the effect of all stressors, individually, cumulatively, and/ or sequentially, on the type and rate of these mechanisms. L&DS has been developed as a novel and robust approach to reliability of PV systems with diagnostic and predictive capabilities. It brings both physical and statistical models and laboratory-based studies and real-world observations together for a pathway of degradation mechanisms. Data-driven epidemiological studies help understand the complex interactions between stressors and degradation responses and the temporal evolution of degradation mechanisms. This new approach will introduce new angles in degradation science and reliability of PV modules.

Acknowledgments The case studies presented in Section 4.2 were performed at the SDLE Research Center (funded through Ohio Third Frontier, Wright Project Program Award Tech 12-004) at Case Western Reserve University.

References [1] J.E. Pickett, Hydrolysis kinetics and lifetime prediction for polycarbonate and polyesters in solar energy applications, in: C.C. White, J. Martin, J.T. Chapin (Eds.), Service Life Prediction of Exterior Plastics, Springer International Publishing, 2015, pp. 41e58. ISBN 978-3-319-06033-0 978-3-319-06034-7, http:// link.springer.com/chapter/10.1007/978-3-31906034-7_3. [2] T. Surek, Progress in U.S. photovoltaics: looking back 30 years and looking ahead 20, in:

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Proceedings of 3rd World Conference on Photovoltaic Energy Conversion, 2003, vol. 3, 2003, pp. 2507e2512, vol. 3. [3] N.G. Dhere, Reliability of PV modules and balance-of-system components, in: Conference Record of the Thirty-First IEEE Photovoltaic Specialists Conference, 2005, 2005, pp. 1570e1576, https://doi.org/10.1109/PVSC. 2005.1488445. [4] A. Luque, Will we exceed 50% efficiency in photovoltaics? Journal of Applied Physics 110 (3) (2011) 031301, https://doi.org/10.1063/ 1.3600702. http://aip.scitation.org/doi/abs/10. 1063/1.3600702. [5] E.D. Dunlop, D. Halton, H.A. Ossenbrink, 20 years of life and more: where is the end of life of a PV module?, in: Conference Record of the Thirty-First IEEE Photovoltaic Specialists Conference, 2005, 2005, pp. 1593e1596, https://doi.org/10.1109/PVSC.2005.1488449. [6] J.H. Wohlgemuth, D.W. Cunningham, P. Monus, J. Miller, A. Nguyen, Long term reliability of photovoltaic modules, in: 2006 IEEE 4th World Conference on Photovoltaic Energy Conference, vol. 2, 2006, pp. 2050e2053, https://doi.org/ 10.1109/WCPEC.2006.279905, vol. 2. [7] C. Bryant, N.R. Wheeler, F. Rubel, R.H. French, Kgc: Koeppen-Geiger Climatic Zones, 2017. [8] F. Rubel, M. Kottek, Observed and projected climate shifts 1901e2100 depicted by world maps of the Koppen-Geiger climate classification, Meteorologische Zeitschrift 19 (2) (2010) 135e141, https://doi.org/10.1127/0941-2948/ 2010/0430. [9] G.A. Klutke, P.C. Kiessler, M.A. Wortman, A critical look at the bathtub curve, IEEE Transactions on Reliability 52 (1) (2003) 125e129, https://doi.org/10.1109/TR.2002. 804492. [10] E.S. Kumar, B. Sarkar, Reliability and vulnerability analysis of solar photovoltaic modules and systems, in: 2013 Annual International Conference on Emerging Research Areas and 2013 International Conference on Microelectronics, Communications and Renewable Energy, 2013, pp. 1e7. https://doi.org/10.1109/ AICERA-ICMiCR.2013.6576028. [11] H.M. Hubbard, Photovoltaics today and tomorrow, Science 244 (4902) (1989) 297e304. http://www.jstor.org/stable/1703250.

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

[12] J.C. Schaefer, Review of photovoltaic power plant performance and economics, IEEE Transactions on Energy Conversion 5 (2) (1990) 232e238. https://doi.org/10.1109/60.107215. [13] A.B. Maish, C. Atcitty, S. Hester, D. Greenberg, D. Osborn, D. Collier, et al., Photovoltaic system reliability, in: Conference Record of the Twenty Sixth IEEE Photovoltaic Specialists Conference e 1997, 1997, pp. 1049e1054. https://doi.org/10.1109/PVSC.1997.654269. [14] D.L. King, M.A. Quintana, J.A. Kratochvil, D.E. Ellibee, B.R. Hansen, Photovoltaic module performance and durability following longterm field exposure, Progress in Photovoltaics: Research and Applications 8 (2) (2000) 241e256. https://doi.org/10.1002/(SICI)1099159X(200003/04)8:2¡241::AID-PIP290¿3.0. CO;2-D. http://onlinelibrary.wiley.com/doi/10. 1002/(SICI)1099-159X(200003/04)8:23.0.CO;2-D/abstract. [15] C.R. Osterwald, A. Anderberg, S. Rummel, L. Ottoson, Degradation analysis of weathered crystalline-silicon PV modules, in: Conference Record of the Twenty-Ninth IEEE Photovoltaic Specialists Conference, 2002, 2002, pp. 1392e 1395. https://doi.org/10.1109/PVSC.2002.119 0869. [16] M. Quintana, D. King, T. McMahon, C. Osterwald, Commonly observed degradation in field-aged photovoltaic modules, in: Conference Record of the Twenty-Ninth IEEE Photovoltaic Specialists Conference, 2002, IEEE, New Orleans, LA, USA, 2002, pp. 1436e1439. https://doi.org/10.1109/PVSC. 2002.1190879. ISBN 978-0-7803-7471-3. [17] J.E. Granata, W.E. Boyson, J.A. Kratochvil, M.A. Quintana, Long-term performance and reliability assessment of 8 PV arrays at Sandia National Laboratories, in: 2009 34th IEEE Photovoltaic Specialists Conference, PVSC, 2009, pp. 001486e001491. https://doi.org/10. 1109/PVSC.2009.5411336. [18] W. Herrmann, N. Bogdanski, F. Reil, M. Ko¨hl, K.A. Weiss, M. Assmus, et al., PV module degradation caused by thermomechanical stress: real impacts of outdoor weathering versus accelerated testing in the laboratory, in: Reliability of Photovoltaic Cells, Modules, Components, and Systems III, vol. 7773, International Society for Optics and Photonics,

71

[19]

[20]

[21]

[22]

[23]

[24]

[25]

2010, p. 77730I. https://doi.org/10.1117/12. 859809, vol. 7773. J. Oh, G. TamizhMani, Temperature testing and analysis of PV modules PER ANSI/UL 1703 and IEC 61730 standards, in: 2010 35th IEEE Photovoltaic Specialists Conference, PVSC, 2010, pp. 000984e000988. https://doi.org/10. 1109/PVSC.2010.5614569. W. Herrmann, N. Bogdanski, Outdoor weathering of PV modules - effects of various climates and comparison with accelerated laboratory testing, in: Proceedings of IEEE PVSC, IEEE, 2011, pp. 002305e002311. https://doi. org/10.1109/PVSC.2011.6186415. ISBN 9781-4244-9965-6 978-1-4244-9966-3 978-14244-9964-9. M.A. Munoz, M.C. Alonso-Garca, N. Vela, F. Chenlo, Early degradation of silicon PV modules and guaranty conditions, Solar Energy 85 (9) (2011) 2264e2274. https://doi.org/10. 1016/j.solener.2011.06.011. http://www.science direct.com/science/article/pii/S0038092X1100 2143. C.E. Chamberlin, M.A. Rocheleau, M.W. Marshall, A.M. Reis, N.T. Coleman, P.A. Lehman, Comparison of PV module performance before and after 11 and 20 years of field exposure, in: 2011 37th IEEE Photovoltaic Specialists Conference, 2011, pp. 000101e 000105. https://doi.org/10.1109/PVSC.2011. 6185854. P. Snchez-Friera, M. Piliougine, J. Pelez, J. Carretero, M. Sidrach de Cardona, Analysis of degradation mechanisms of crystalline silicon PV modules after 12 years of operation in Southern Europe, Progress in Photovoltaics: Research and Applications 19 (6) (2011) 658e666. https://doi.org/10.1002/pip.1083. http://onlinelibrary.wiley.com/doi/10.1002/pip. 1083/abstract. E.S. Kopp, V.P. Lonij, A.E. Brooks, P.L. Hidalgo-Gonzalez, A.D. Cronin, IV curves and visual inspection of 250 PV modules deployed over 2 years in tucson, in: 2012 38th IEEE Photovoltaic Specialists Conference, 2012, pp. 003166e003171. https://doi.org/10. 1109/PVSC.2012.6318251. K. Matsuda, T. Watanabe, K. Sakaguchi, M. Yoshikawa, T. Doi, A. Masuda, Microscopic degradation mechanisms in silicon photovoltaic

72

[26]

[27]

[28]

[29]

[30]

[31]

[32]

D URABILITY AND R ELIABILITY

OF

P OLYMERS

module under long-term environmental exposure, Japanese Journal of Applied Physics 51 (10S) (2012) 10NF07. https://doi.org/10.1143/ JJAP.51.10NF07. http://iopscience.iop.org/ article/10.1143/JJAP.51.10NF07/meta. T.M. Walsh, Z. Xiong, Y.S. Khoo, A.A. Tay, A.G. Aberle, Singapore modules e optimised PV modules for the tropics, Energy Procedia 15 (2012) 388e395. https://doi.org/10.1016/j. egypro.2012.02.047. http://www.sciencedirect. com/science/article/pii/S1876610212003839. D. Wu, J. Zhu, T.R. Betts, R. Gottschalg, PV module degradation mechanisms under different environmental stress factors, in: Proceedings of the 8th Photovoltaic Science Application and Tech- Nology Conference (PVSAT-8), The Solar Energy Society, 2012. https://dspace.lboro.ac.uk/dspace-jspui/handle/ 2134/13517. D.C. Jordan, S.R. Kurtz, Photovoltaic degradation ratesan analytical review, Progress in Photovoltaics: Research and Applications 21 (1) (2013) 12e29. https://doi.org/10.1002/ pip.1182. http://onlinelibrary.wiley.com/doi/ 10.1002/pip.1182/abstract. A. Ndiaye, A. Charki, A. Kobi, C.M.F. Kb, P.A. Ndiaye, V. Sambou, Degradations of silicon photovoltaic modules: a literature review, Solar Energy 96 (Supplement C) (2013) 140e151. https://doi.org/10.1016/j.solener. 2013.07.005. http://www.sciencedirect.com/ science/article/pii/S0038092X13002703. V. Sharma, S.S. Chandel, Performance and degradation analysis for long term reliability of solar photovoltaic systems: a review, Renewable and Sustainable Energy Reviews 27 (Supplement C) (2013) 753e767. https://doi.org/10. 1016/j.rser.2013.07.046. http://www.science direct.com/science/article/pii/S136403211300 4917. W. Gambogi, Y. Heta, K. Hashimoto, J. Kopchick, T. Felder, S. MacMaster, et al., A comparison of key PV backsheet and module performance from fielded module exposures and accelerated tests, IEEE Journal of Photovoltaics 4 (3) (2014) 935e941. https://doi.org/ 10.1109/JPHOTOV.2014.2305472. http://iee explore.ieee.org/document/6782472/. M. Kontges, S. Kurtz, C. Packard, U. Jahn, K. Berger, K. Kato, et al., Performance and

AND

OTHER M ATERIALS

[33]

[34]

[35]

[36]

[37]

[38]

IN

P HOTOVOLTAIC M ODULES

Reliability of Photovoltaic Systems: Review of Failures of PV Modules, International Energy Agency, 2014. Tech. Rep. IEA-PVPS Task13-01:2014, http://iea-pvps.org/index. php?id¼275. M. Koehl, N. Bogdanski, K.A. Wei, I. Drr, S. Hoffmann, C. Peike, PV module degradation in the field and in the lab - how does it fit together?, in: 29th European Photovoltaic Solar Energy Conference and Exhibition, 2014, pp. 3110e3114. https://doi.org/10.4229/ EUPVSEC20142014-5BV.4.1, in: http://www. eupvsec-proceedings.com/proceedings? paper¼29385. G. Makrides, J. Bratcher, J. Pratt, G. Georghiou, Outdoor Performance and Reliability Evaluation of Different Crystalline Silicon Photovoltaic Modules with Innovative Packaging Components, 2014. https://www.researchgate. net/publication/295857922_Outdoor_Perform ance_And_Reliability_Evaluation_Of_Different_ Crystalline_Silicon_Photovoltaic_Modules_ With_Innovative_Packaging_Components. G. Makrides, I. Koumparou, J. Bratcher, J. Pratt, G. Georghiou, Thermal Classification Modelling and Energy Yield Performance of Different Crystalline Silicon Photovoltaic Modules with Innovative Packaging Components, 2016. https://www.researchgate.net/publication/3060 17013_THERMAL_Classification_Modelling_ And_Energy_Yield_Performance_Of_Different_ Crystalline_Silicon_Photovoltaic_Modules_ With_Innovative_Packaging_Components. A. Pozza, T. Sample, Crystalline silicon PV module degradation after 20 years of field exposure studied by electrical tests, electroluminescence, and LBIC, Progress in Photovoltaics: Research and Applications 24 (3) (2016) 368e378. https://doi.org/10.1002/pip.2717. D.C. Jordan, S.R. Kurtz, K. VanSant, J. Newmiller, Compendium of photovoltaic degradation rates, Progress in Photovoltaics: Research and Applications 24 (7) (2016) 978e989. https://doi.org/10.1002/pip.2744. http://onlinelibrary.wiley.com/doi/10.1002/ pip.2744/abstract. V. Sharma, S.S. Chandel, A novel study for determining early life degradation of multicrystalline-silicon photovoltaic modules observed in western Himalayan Indian climatic

3: D EGRADATION S CIENCE

[39]

[40]

[41]

[42]

[43]

[44]

[45]

AND

PATHWAYS

IN

PV S YSTEMS

conditions, Solar Energy 134 (Supplement C) (2016) 32e44. https://doi.org/10.1016/j. solener.2016.04.023. http://www.sciencedirect. com/science/article/pii/S0038092X16300561. D.A. Quansah, M.S. Adaramola, G. Takyi, I.A. Edwin, Reliability and degradation of solar PV ModulesCase study of 19-year-old polycrystalline modules in Ghana, Technologies 5 (2) (2017) 22. https://doi.org/10.3390/ technologies5020022. http://www.mdpi.com/ 2227-7080/5/2/22. S. Tatapudi, C. Libby, C. Raupp, D. Srinivasan, J. Kuitche, B. Bicer, et al., Defect and safety inspection of 6 PV technologies from 56,000 modules representing 257,000 modules in 4 climatic regions of the United States, in: 2016 IEEE 43rd Photovoltaic Specialists Conference, PVSC, 2016, pp. 1747e1751. https://doi. org/10.1109/PVSC.2016.7749923. D.C. Jordan, T.J. Silverman, J.H. Wohlgemuth, S.R. Kurtz, K.T. VanSant, Photovoltaic failure and degradation modes, Progress in Photovoltaics: Research and Applications 25 (4) (2017) 318e326. https://doi.org/10.1002/pip.2866. http://onlinelibrary.wiley.com/doi/10.1002/pip. 2866/abstract. D.C. Jordan, T.J. Silverman, B. Sekulic, S.R. Kurtz, PV degradation curves: nonlinearities and failure modes, Progress in Photovoltaics: Research and Applications 25 (7) (2017) 583e591. https://doi.org/10.1002/pip. 2835. http://onlinelibrary.wiley.com/doi/10. 1002/pip.2835/abstract. M. Kontges, G. Oreski, U. Jahn, M. Herz, P. Hacke, K.A. Weiss, et al., Assessment of Photovoltaic Module Failures in the Field, International Energy Agency, 2017. Tech. Rep. IEA-PVPS T13-09:2017, http://iea-pvps.org/ index.php?id¼275. A. Fairbrother, M. Boyd, Y. Lyu, J. Avenet, P. Illich, Y. Wang, et al., Differential degradation patterns of photovoltaic backsheets at the array level, Solar Energy 163 (2018) 62e69. http://www.sciencedirect.com/science/article/ pii/S0038092X18300938. https://doi.org/10. 1016/j.solener.2018.01.072. M.P. Murray, L.S. Bruckman, R.H. French, Durability of materials in a stress-response framework: acrylic materials for photovoltaic systems, in: MRS Online Proceedings Library,

73

[46]

[47]

[48]

[49]

[50]

[51]

[52]

vol. 1391, 2012. https://doi.org/10.1557/opl. 2012.1241 vol. 1391, https://doi.org/10.1557/ opl.2012.1241. M.P. Murray, L.S. Bruckman, R.H. French, Photodegradation in a stress and response framework: poly(methyl methacrylate) for solar mirrors and lens, Journal of Photonics for Energy 2 (1) (2012). https://doi.org/10.1117/1. JPE.2.022004, 022004e022004, https://doi. org/10.1117/1.JPE.2.022004. L.S. Bruckman, N.R. Wheeler, J. Ma, E. Wang, C.K. Wang, I. Chou, et al., Statistical and domain analytics applied to PV module lifetime and degradation science, IEEE Access 1 (2013) 384e403. https://doi.org/10.1109/ACCESS. 2013.2267611. L.S. Bruckman, N.R. Wheeler, I.V. Kidd, J. Sun, R.H. French, Photovoltaic lifetime and degradation science statistical pathway development: acrylic degradation, in: N.G. Dhere, J.H. Wohlgemuth, K.W. Lynn (Eds.), Proc. SPIE 8825, Reliability of Photovoltaic Cells, Modules, Components, and Systems VI, 2013, p. 88250D. https://doi.org/10.1117/12.2024717, in: http://proceedings.spiedigitallibrary.org/ proceeding.aspx?doi¼10.1117/12.2024717. N.R. Wheeler, L. Bruckman, J. Ma, E. Wang, C. Wang, I. Chou, et al., Statistical and domain analytics for informed study protocols, in: 2013 IEEE Energytech, 2013, pp. 1e7. https://doi. org/10.1109/EnergyTech.2013.6645354. N.R. Wheeler, Y. Xu, A. Gok, I.V. Kidd, L.S. Bruckman, J. Sun, et al., Data science study protocols for investigating lifetime and degradation of PV technology systems, in: IEEE PVSC 40, Denver, Colorado, 2014, 00000. N.R. Wheeler, A. Gok, T.J. Peshek, L.S. Bruckman, N. Goel, D. Zabiyaka, et al., A data science approach to understanding photovoltaic module degradation, in: Proc. of SPIE, vol. 9563, SPIE, 2015. https://doi.org/10. 1117/12.2209204, pp. 95630Le95630Le9563 6L, https://doi.org/10.1117/12.2209204. A. Gok, C.L. Fagerholm, D.A. Gordon, L.S. Bruckman, R.H. French, Degradation of poly(ethylene-terephthalate) under accelerated weathering exposures, in: Photovoltaic Specialist Conference (PVSC), 2015 IEEE 42nd, 2015, pp. 1e6. https://doi.org/10.1109/ PVSC.2015.7355601.

74

D URABILITY AND R ELIABILITY

OF

P OLYMERS

[53] R.H. French, R. Podgornik, T.J. Peshek, L.S. Bruckman, Y. Xu, N.R. Wheeler, et al., Degradation science: mesoscopic evolution and temporal analytics of photovoltaic energy materials, Current Opinion in Solid State & Materials Science 19 (4) (2015) 212e226. https:// doi.org/10.1016/j.cossms.2014.12.008. http:// www.sciencedirect.com/science/article/pii/S13 59028614000989. [54] L.S. Bruckman, R.H. French, Y. Wang, M.D. Kempe, A.A. Lefebvre, X. Gu, et al., Cross-correlation of backsheet degradation between real-world exposed modules and accelerated exposures of backsheet materials (Conference Presentation), in: N.G. Dhere, J.H. Wohlgemuth, K. Sakurai (Eds.), SPIE Proceedings, 2016, p. 993805. https://doi.org/ 10.1117/12.2238162, in: http://proceedings.spi edigitallibrary.org/proceeding.aspx?doi¼10.11 17/12.2238162. [55] Y. Wang, A. Gok, C.L. Fagerholm, L.S. Bruckman, R.H. French, Development of Predictive and Semi-gSEM Models of Backsheet Degradation under Multifactor Accelerated Weathering Exposure, Society of Plastics Engineers, Indianapolis, Indiana, USA, 2016, pp. 575e580. ISBN 978-0-692-71961-9, www. 4spe.org. [56] A.G. Klinke, A. Gok, S.I. Ifeanyi, R.H. French, L.S. Bruckman, Degradation of Photovoltaic Backsheet Materials under Multi-Factor Accelerated UV Light Exposures, vol. 10370, International Society for Optics and Photonics, 2017, p. 1037004. https://doi.org/10.1117/12. 2274108, in: https://www.spiedigitallibrary.org/ conference-proceedings-of-spie/10370/103700 4/Degradation-of-photovoltaic-backsheet-mate rials-under-multi-factor-accelerated-UV/10.11 17/12.2274108.short. [57] A. Gok, D.K. Ngendahimana, C.L. Fagerholm, R.H. French, J. Sun, L.S. Bruckman, Predictive models of poly(ethylene-terephthalate) film degradation under multi-factor accelerated weathering exposures, PLoS One 12 (5) (2017). https://doi.org/10.1371/journal.pone.0177614. http://dx.plos.org/10.1371/journal.pone.01776 14. [58] A. Gok, C.L. Fagerholm, R.H. French, L.S. Bruckman, Temporal evolution and pathway models of poly(ethylene-terephthalate)

AND

OTHER M ATERIALS

[59]

[60]

[61]

[62]

[63]

[64]

[65] [66]

[67]

IN

P HOTOVOLTAIC M ODULES

degradation under multi-factor accelerated weathering exposures, PLoS One 14 (2) (2019) e0212258. https://doi.org/10.1371/journal. pone.0212258. A. Gok, D.A. Gordon, D.M. Burns, S.P. Fowler, R.H. French, L.S. Bruckman, Reciprocity and spectral effects of the degradation of poly(ethylene-terephthalate) under accelerated weathering exposures, Journal of Applied Polymer Science 136 (22) (2019). https://doi. org/10.1002/app.47589. D.A. Gordon, W.H. Huang, D.M. Burns, R.H. French, L.S. Bruckman, Multivariate multiple regression models of poly(ethyleneterephthalate) film degradation under outdoor and multi-stressor accelerated weathering exposures, PLoS One 13 (12) (2018) e0209016. https://doi.org/10.1371/journal.pone.0209016. D.A. Gordon, Z. Zhan, L.S. Bruckman, Characterizing the weathering induced degradation of Poly(ethylene-terephthalate) using PARAFAC modeling of fluorescence spectra, Polymer Degradation and Stability 161 (2019) 85e94. https://doi.org/10.1016/j.polymdegrad stab.2019.01.006. K.B. Natasha, S. Guo, Structural equation modeling, in: Social Work Research Methods, first ed., Oxford University Press, Oxford; New York, 2011. ISBN 978-0-19-536762-1. R.B. Kline, Principles and Practice of Structural Equation Modeling, Guilford Press, 2011. ISBN 978-1-60623-876-9. 21203. W.H. Huang, N. Wheeler, A. Klinke, Y. Xu, W. Du, A.K. Verma, et al., netSEM: Network Structural Equation Modeling, 2018. Deleted in review. K.J. Lewis, Encapsulant material requirements for photovoltaic modules, in: Polymers in Solar Energy Utilization, ACS Symposium Series, ACS Symposium Series, vol. 220, American Chemical Society, 1983, pp. 367e385. https:// doi.org/10.1021/bk-1983-0220.ch023. ISBN 978-0-8412-0776-9, https://doi.org/10.1021/ bk-1983-0220.ch023. M. Kempe, Overview of scientific issues involved in selection of polymers for PV applications, in: 2011 37th IEEE Photovoltaic Specialists Conference, 2011, pp. 000085e 000090. https://doi.org/10.1109/PVSC.2011. 6185851.

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

[68] S.H. Schulze, M. Pander, S. Dietrich, M. Ebert, Encapsulation polymers e a key issue in module reliability, Photovoltaics International 11 (2011). https://www.pv-tech.org/technicalpapers/encapsulation-polymers-a-key-issue-inmodule-reliability. [69] C. Peike, I. Durr, I. Hadrich, K. Weiss, Overview of PV module encapsulation materials, Photovoltaics International 19 (2013). https:// www.pv-tech.org/technical-papers/overview-ofpv-module-encapsulation-materials. [70] G. Griffini, S. Turri, Polymeric materials for long-term durability of photovoltaic systems, Journal of Applied Polymer Science 133 (11) (2016). https://doi.org/10.1002/app.43080 n/ aen/a, http://doi.wiley.com/10.1002/app.43080. [71] H.Y. Li, Y. Luo, C. Ballif, L.E. Perret-Aebi, Effect of cooling press on the optical transmission through photovoltaic encapsulants, Polymer e Plastics Technology and Engineering 54 (4) (2015) 416e424. https://doi.org/10. 1080/03602559.2014.958778. https://doi.org/ 10.1080/03602559.2014.958778. [72] K.J. Geretschlager, G.M. Wallner, J. Fischer, Structure and basic properties of photovoltaic module backsheet films, Solar Energy Materials and Solar Cells 144 (Supplement C) (2016) 451e456. https://doi.org/10.1016/j.solmat.2015. 09.060. http://www.sciencedirect.com/science/ article/pii/S0927024815005024. [73] M.D. Kempe, P. Thapa, Low cost, single layer replacement for the back-sheet and encapsulant layers, in: N.G. Dhere (Ed.), Reliability of Photovoltaic Cells, Modules, Components, and Systems, vol. 7048, Spie-Int Soc Optical Engineering, Bellingham, 2008. ISBN 978-08194-7268-7. 00003 WOS:000262507800013. [74] U. Hoyer, T. Swonke, Diffusion of moisture and impact of UV irradiance in photovoltaic encapsulants, in: 24th European Photovoltaic Solar Energy Conference, 21e25 September 2009, Hamburg, Germany, 2009, pp. 3373e3376. https://doi.org/10.4229/24thEUPVSEC2009-4 AV.3.23, in: http://www.eupvsec-proceedings. com/proceedings?paper¼4642. [75] T. Swonke, R. Auer, Impact of Moisture on PV Module Encapsulants, 2009, p. 74120A. https:// doi.org/10.1117/12.825943, in: http://procee dings.spiedigitallibrary.org/proceeding.aspx? doi¼10.1117/12.825943.

75

[76] B. Ketola, A. Norris, The role of encapsulant moisture permeability in the durability of solar photovoltaic modules, in: Proc 25th EU PVSEC, 2010, pp. 4098e4102. http://www. dowcorning.com/content/publishedlit/06-1066 -01.pdf. [77] G. Pinter, G. Oreski, Comparative study of the temperature dependent delamination behavior of four solar cell encapsulants to glass and backsheet-laminate, in: 26th European Photovoltaic Solar Energy Conference and Exhibition, 2011, pp. 3305e3309. https://doi.org/10. 4229/26thEUPVSEC2011-4AV.1.4, in: http:// www.eupvsec-proceedings.com/proceedings? paper¼11073. [78] S. Wang, X. Sun, Y. Wu, Y. Huang, N. Shepherd, G. Beaucarne, Double-glass PV modules with silicone encapsulation, Photovoltaics International 31 (2016). https://www.pv-tech.org/ technical-papers/double-glass-pv-modules-withsilicone-encapsulation. [79] M.D. Kempe, D.C. Miller, J.H. Wohlgemuth, S.R. Kurtz, J.M. Moseley, D.L. Nobles, et al., Multi angle laser light scattering evaluation of field exposed thermoplastic photovoltaic encapsulant materials, Energy Science & Engineering 4 (1) (2016) 40e51. https://doi.org/ 10.1002/ese3.106. http://doi.wiley.com/10. 1002/ese3.106. [80] M. DeBergalis, Fluoropolymer films in the photovoltaic industry, Journal of Fluorine Chemistry 125 (8) (2004) 1255e1257. https:// doi.org/10.1016/j.jfluchem.2004.05.013. http:// www.sciencedirect.com/science/article/pii/S00 22113904001472. [81] S.B. Levy, Bio-based Backsheet, 2008, p. 70480C. https://doi.org/10.1117/12.792646, in: http://proceedings.spiedigitallibrary.org/pro ceeding.aspx?doi¼10.1117/12.792646. [82] Biobased PV backsheet, in: S. Levy (Ed.), 2011 37th IEEE Photovoltaic Specialists Conference, PVSC, 2011, pp. 003185e003188. https://doi. org/10.1109/PVSC.2011.6186617. [83] T. Gorter, A. Reinders, A comparison of 15 polymers for application in photovoltaic modules in PV-powered boats, Applied Energy 92 (2012) 286e297. https://doi.org/10.1016/j. apenergy.2011.10.042. http://linkinghub.elsevi er.com/retrieve/pii/S0306261911007033.

76

D URABILITY AND R ELIABILITY

OF

P OLYMERS

[84] M. Knausz, G. Oreski, G.C. Eder, Y. Voronko, B. Duscher, T. Koch, et al., Degradation of photovoltaic backsheets: comparison of the aging induced changes on module and component level, Journal of Applied Polymer Science 132 (24) (2015). https://doi.org/10.1002/app. 42093 n/aen/a, http://doi.wiley.com/10.1002/ app.42093. [85] K. Kim, M. Yoo, K. Ahn, J. Kim, Thermal and mechanical properties of AlN/BN-filled PVDF composite for solar cell backsheet application, Ceramics International 41 (1, Part A) (2015) 179e187. https://doi.org/10.1016/j.ceramint. 2014.08.056. http://www.sciencedirect.com/ science/article/pii/S0272884214012930. [86] K.J. Geretschlager, G.M. Wallner, I. Hintersteiner, W. Buchberger, Damp heat aging behavior of a polyamide-based backsheet for photovoltaic modules, Journal of Solar Energy Engineering 138 (4) (2016) 041003. https://doi.org/10.1115/1.4032977. http://solar energyengineering.asmedigitalcollection.asme. org/article.aspx?doi¼10.1115/1.4032977. [87] G.M. Wallner, K.J. Geretschlger, I. Hintersteiner, W. Buchberger, Characterization of polyoxymethylene for backsheets of photovoltaic modules, Journal of Plastic Film and Sheeting 33 (4) (2016) 345e360. https:// doi.org/10.1177/8756087916666624, 875608 7916666624, https://doi.org/10.1177/8756087 916666624. [88] G. Barber, G. Jorgensen, K. Terwilliger, S. Glick, J. Pern, T. McMahon, New Barrier Coating Materials for PV Module Backsheets, IEEE, 2002, pp. 1541e1544. https://doi.org/10. 1109/PVSC.2002.1190906. ISBN 978-0-78037471-3, http://ieeexplore.ieee.org/document/ 1190906/. [89] G. Jorgensen, K. Terwilliger, S. Glick, J. Pern, T. McMahon, Materials Testing for PV Module Encapsulation, 2003. Denver, Colorado, https:// www.nrel.gov/docs/fy03osti/33578.pdf. [90] M.D. Kempe, Control of moisture ingress into photovoltaic modules, in: Conference Record of the Thirty-First IEEE Photovoltaic Specialists Conference, 2005, 2005, pp. 503e506. https:// doi.org/10.1109/PVSC.2005.1488180.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

[91] M.D. Kempe, Modeling of rates of moisture ingress into photovoltaic modules, Solar Energy Materials and Solar Cells 90 (16) (2006) 2720e2738. https://doi.org/10.1016/j.solmat. 2006.04.002. http://www.sciencedirect.com/ science/article/pii/S0927024806001632. [92] M. Koehl, O. Angeles-Palacios, D. Philipp, K.A. Wei, Polymer films in photovoltaic modules: analysis and modeling of permeation processes, in: J.W. Martin, R.A. Ryntz, J. Chin, R.A. Dickie (Eds.), Service Life Prediction of Polymeric Materials, Springer US, Boston, MA, 2009, pp. 361e371. https://doi.org/10. 1007/978-0-387-84876-1_24. ISBN 978-0387-84875-4 978-0-387-84876-1, http://link. springer.com/10.1007/978-0-387-84876-1_24. [93] P. Hulsmann, D. Philipp, M. Khl, Measuring temperature-dependent water vapor and gas permeation through high barrier films, Review of Scientific Instruments 80 (11) (2009) 113901. https://doi.org/10.1063/1.3250866. [94] P. Hulsmann, C. Peike, M. Blml, P. Schmid, K.A. Wei, M. Khl, Impact of permeation properties and back sheet/encapsulation interactions on the reliability of PV modules, ASME (2011) 1767e1771. https://doi.org/10.1115/IMECE20 11-65723. ISBN 978-0-7918-5490-7, http:// proceedings.asmedigitalcollection.asme.org/ proceeding.aspx?articleid¼1643561. [95] P. Hlsmann, K.A. Wei, M. Khl, Temperaturedependent water vapour and oxygen permeation through different polymeric materials used in photovoltaic-modules, Progress in Photovoltaics: Research and Applications 22 (4) (2014) 415e421. https://doi.org/10.1002/ pip.2273. http://onlinelibrary.wiley.com/doi/ 10.1002/pip.2273/abstract. [96] Y. Zhang, J. Xu, J. Mao, J. Tao, H. Shen, Y. Chen, et al., Long-term reliability of silicon wafer-based traditional backsheet modules and double glass modules, RSC Advances 5 (81) (2015) 65768e65774. https://doi.org/10.1039/ C5RA11224A. http://pubs.rsc.org/en/content/ articlelanding/2015/ra/c5ra11224a. [97] Y. Voronko, G.C. Eder, M. Knausz, G. Oreski, T. Koch, K.A. Berger, Correlation of the loss in photovoltaic module performance with the

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

ageing behaviour of the backsheets used: influence of backsheet on photovoltaic module performance, Progress in Photovoltaics: Research and Applications 23 (11) (2015) 1501e1515. https://doi.org/10.1002/pip.2580. http://doi.wiley.com/10.1002/pip.2580. [98] P. Hlsmann, G.M. Wallner, Permeation of water vapour through polyethylene terephthalate (PET) films for back-sheets of photovoltaic modules, Polymer Testing 58 (2017) 153e158. https://doi.org/10.1016/j.polymertesting.2016. 11.028. http://www.sciencedirect.com/science/ article/pii/S0142941816305396. [99] G. Oreski, A. Mihaljevic, Y. Voronko, G.C. Eder, Acetic acid permeation through photovoltaic backsheets: influence of the composition on the permeation rate, Polymer Testing 60 (Supplement C) (2017) 374e380. https://doi.org/10.1016/j.polymertesting.2017. 04.025. http://www.sciencedirect.com/science/ article/pii/S014294181730209X. [100] G. Jorgensen, K. Terwilliger, M. Kempe, J. Pern, S. Glick, J. del Cueto, et al., Packaging Materials and Design for Improved PV Module Reliability, 2004. [101] G.J. Jorgensen, K.M. Terwilliger, J.A. DelCueto, S.H. Glick, M.D. Kempe, J.W. Pankow, et al., Moisture transport, adhesion, and corrosion protection of PV module packaging materials, Solar Energy Materials and Solar Cells 90 (16) (2006) 2739e2775. https://doi.org/10.1016/j.solmat.2006.04.003. http://www.sciencedirect.com/science/article/ pii/S0927024806001644. [102] J.H. Wohlgemuth, P. Hacke, N. Bosco, D.C. Miller, M.D. Kempe, S.R. Kurtz, Assessing the Causes of Encapsulant Delamination in PV Modules, IEEE, 2016, pp. 0248e0254. https://doi.org/10.1109/PVSC.2016.7749589. ISBN 978-1-5090-2724-8, http://ieeexplore. ieee.org/document/7749589/. [103] C. Ballif, R.F.M. Lange, T. Turlings, Y. LuoHoffmann, G. Rder, R. Thron, et al., Insights on EVA lamination process: where do the bubbles come from?, in: 25th European Photovoltaic Solar Energy Conference and Exhibition/5th World Conference on Photovoltaic Energy Conversion, 6e10 September 2010, Valencia, Spain, 2010, pp. 4036e4038. https://doi.org/10.4229/25thEUPVSEC2010-

77

[104]

[105]

[106]

[107]

[108]

[109]

[110]

4AV.3.24, in: http://www.eupvsec-proceedi ngs.com/proceedings?paper¼5844. W.J. Gambogi, Comparative performance of backsheets for photovoltaic modules, in: 25th European Photovoltaic Solar Energy Conference and Exhibition/5th World Conference on Photovoltaic Energy Conversion, 6e10 September 2010, Valencia, Spain, 2010, pp. 4079e4083. https://doi.org/10.4229/25th EUPVSEC2010-4AV.3.39, in: http://www. eupvsec-proceedings.com/proceedings?paper ¼8122. K. Agroui, M. Jaunich, A.H. Arab, Analysis techniques of polymeric encapsulant materials for photovoltaic modules: situation and perspectives, Energy Procedia 93 (Supplement C) (2016) 203e210. https://doi.org/10.1016/j. egypro.2016.07.171. http://www.sciencedirect. com/science/article/pii/S1876610216305987. A.W. Czanderna, F.J. Pern, Encapsulation of PV modules using ethylene vinyl acetate copolymer as a pottant: a critical review, Solar Energy Materials and Solar Cells 43 (2) (1996) 101e181. https://doi.org/10.1016/0927-0248 (95)00150-6. http://www.sciencedirect.com/ science/article/pii/0927024895001506. F.J. Pern, S.H. Glick, Thermal processing of EVA encapsulants and effects of formulation additives [for solar cells], in: Conference Record of the Twenty Fifth IEEE Photovoltaic Specialists Conference e 1996, 1996, pp. 1251e1254. https://doi.org/10.1109/PVSC. 1996.564359. G. Oreski, G. Wallner, Delamination behaviour of multi-layer films for PVencapsulation, Solar Energy Materials and Solar Cells 89 (2e3) (2005) 139e151. https://doi.org/10.1016/j. solmat.2005.02.009. http://linkinghub.elsevier. com/retrieve/pii/S0927024805000668. H.Y. Li, R. Thron, G. Rder, T. Turlings, Y. Luo, R. Lange, et al., Insights into the encapsulation process of photovoltaic modules: GC-MS analysis on the curing step of Poly (ethyleneco-vinyl acetate) (EVA) encapsulant, Polymers and Polymer Composites 20 (8) (2012) 665e672. https://infoscience.epfl.ch/record/ 177325. M. Kontges, M. Denz, U. Giese, T. Alshuth, T. Dring, M. Siegert, et al., Influence of the curing state of ethylene-vinyl acetate on

78

[111]

[112]

[113]

[114]

[115]

[116]

D URABILITY AND R ELIABILITY

OF

P OLYMERS

photovoltaic modules aging, in: 28th European Photovoltaic Solar Energy Conference and Exhibition, 2013, pp. 2832e2837. https://doi. org/10.4229/28thEUPVSEC2013-4CO.9.5, in: http://www.eupvsec-proceedings.com/proceed ings?paper¼22902. M. Kraft, B.S. Chernev, G.C. Eder, A. Rauschenbach, G. Oreski, M. DeBiasio, et al., Post-crosslinking in photovoltaic modules under different conditions, in: 29th European Photovoltaic Solar Energy Conference and Exhibition, 2014, pp. 3133e3137. https://doi. org/10.4229/EUPVSEC20142014-5BV.4.9, in: http://www.eupvsec-proceedings.com/proc eedings?paper¼28812. H.Y. Li, L.E. Perret-Aebi, V. Chapuis, C. Ballif, Y. Luo, The effect of cooling press on the encapsulation properties of crystalline photovoltaic modules: residual stress and adhesion, Progress in Photovoltaics: Research and Applications 23 (2) (2015) 160e169. https://doi. org/10.1002/pip.2409. http://onlinelibrary. wiley.com/doi/10.1002/pip.2409/abstract. H. Li, L. Yun, C. Ballif, L.E. Perret-Aebi, Modeling of voids evolution in the encapsulation process of photovoltaic modules, Polymers and Polymer Composites 23 (6) (2015) 375e388. https://infoscience.epfl.ch/record/ 204752. M. Knausz, G. Oreski, M. Schmidt, P. Guttmann, K. Berger, Y. Voronko, et al., Thermal expansion behavior of solar cell encapsulation materials, Polymer Testing 44 (2015) 160e167. https://doi.org/10.1016/j. polymertesting.2015.04.009. http://www. sciencedirect.com/science/article/pii/S014294 1815000999. K. Lin, Y.H. Lee, W.Y. Huang, G. Chen, Y.W. Kuo, L.K. Wang, et al., Detection of soldering induced damages on crystalline silicon solar modules fabricated by hot-air soldering method, Renewable Energy 83 (Supplement C) (2015) 749e758. https://doi.org/10.1016/j. renene.2015.05.017. http://www.sciencedirect. com/science/article/pii/S0960148115003948. M. Paggi, A. Sapora, An accurate thermoviscoelastic rheological model for ethylene vinyl acetate based on fractional calculus, International Journal of Photoenergy 2015 (2015). https://doi.org/10.1155/2015/252740.

AND

OTHER M ATERIALS

[117]

[118]

[119]

[120]

[121]

[122]

[123]

IN

P HOTOVOLTAIC M ODULES

https://www.hindawi.com/journals/ijp/2015/ 252740/. K. Agroui, G. Collins, G. Oreski, M. Boehning, A. Hadj Arab, D. Ouadjaout, Effect of crosslinking on EVA-based encapsulant properties during photovoltaic module fabrication process, Revue des Energies Renouvelables 18 (2015) 303e314. K. Agroui, G. Collins, F. Giovanni, W. Stark, A comprehensive indoor and outdoor aging of the cross-linked EVA encapsulant material for photovoltaic conversion, Polymer - Plastics Technology and Engineering 54 (7) (2015) 719e729. https://doi.org/10.1080/03602559. 2014.974278. http://www.tandfonline.com/ doi/full/10.1080/03602559.2014.974278. F.D. Novoa, D.C. Miller, R.H. Dauskardt, Adhesion and debonding kinetics of photovoltaic encapsulation in moist environments: adhesion and debonding kinetics of photovoltaic encapsulation, Progress in Photovoltaics: Research and Applications 24 (2) (2016) 183e194. https://doi.org/10.1002/pip.2657. http://doi.wiley.com/10.1002/pip.2657. G. Oreski, A. Rauschenbach, C. Hirschl, M. Kraft, G. Eder, G. Pinter, Crosslinking and post-crosslinking of ethylene vinyl acetate in photovoltaic modules, Journal of Applied Polymer Science 134 (23) (2017). https://doi. org/10.1002/app.44912. http://doi.wiley.com/ 10.1002/app.44912. C. Dechthummarong, B. Wiengmoon, D. Chenvidhya, C. Jivacate, K. Kirtikara, Physical deterioration of encapsulation and electrical insulation properties of PV modules after long-term operation in Thailand, Solar Energy Materials and Solar Cells 94 (9) (2010) 1437e1440. https://doi.org/10.1016/j.solmat. 2010.03.038. http://www.sciencedirect.com/ science/article/pii/S0927024810002011. D. Binesti, N. Chatagnon, J.F. Penneau, A. Plotton, M. Van Iseghem, Failure modes after damp heat tests on PV modules, in: 27th European Photovoltaic Solar Energy Conference and Exhibition, 2012, pp. 3546e3548. https://doi.org/10.4229/27thEUPVSEC20124BV.3.47, in: http://www.eupvsec-proceedi ngs.com/proceedings?paper¼20855. V. Chapuis, F. Galliano, C. Schlumpf, H.Y. Li, C. Ballif, L.E. Perret-Aebi, Methods to

3: D EGRADATION S CIENCE

[124]

[125]

[126]

[127]

[128]

[129]

[130]

AND

PATHWAYS

IN

PV S YSTEMS

Evaluate the Effect of Water Ingress: Towards Ultra-reliable PV Modules, 2013, p. 88250E. https://doi.org/10.1117/12.2028918, in: http:// proceedings.spiedigitallibrary.org/proceeding. aspx?doi¼10.1117/12.2028918. J. Zhu, M. Koehl, S. Hoffmann, K.A. Berger, S. Zamini, I. Bennett, et al., Changes of solar cell parameters during damp-heat exposure: changes of solar cell parameters, Progress in Photovoltaics: Research and Applications 24 (10) (2016) 1346e1358. https://doi.org/10. 1002/pip.2793. http://doi.wiley.com/10.1002/ pip.2793. M.D. Kempe, G.J. Jorgensen, K.M. Terwilliger, T.J. McMahon, C.E. Kennedy, T.T. Borek, Ethylene-vinyl acetate potential problems for photovoltaic packaging, in: 2006 IEEE 4th World Conference on Photovoltaic Energy Conference, vol. 2, 2006, pp. 2160e2163. https://doi.org/10.1109/WCPEC.2006.279933. M.D. Kempe, G.J. Jorgensen, K.M. Terwilliger, T.J. McMahon, C.E. Kennedy, T.T. Borek, Acetic acid production and glass transition concerns with ethylene-vinyl acetate used in photovoltaic devices, Solar Energy Materials and Solar Cells 91 (4) (2007) 315e329. https:// doi.org/10.1016/j.solmat.2006.10.009. http:// www.sciencedirect.com/science/article/pii/S0 927024806004107. T. Shioda, Acetic acid production rate in EVA encapsulant and its influence on performance of PV modules. Presented at the 2nd Atlas/NIST PV Materials Durability Workshop. https:// www.nist.gov/sites/default/files/documents/el/ building_materials/Shioda.pdf, 2013. T.H. Kim, N.C. Park, D.H. Kim, The effect of moisture on the degradation mechanism of multi-crystalline silicon photovoltaic module, Microelectronics Reliability 53 (9) (2013) 1823e1827. https://doi.org/10.1016/j.micro rel.2013.07.047. http://www.sciencedirect. com/science/article/pii/S0026271413002229. A. Kraft, L. Labusch, T. Ensslen, I. Drr, J. Bartsch, M. Glatthaar, et al., Investigation of acetic acid corrosion impact on printed solar cell contacts, IEEE Journal of Photovoltaics 5 (3) (2015) 736e743. https://doi.org/10.1109/ JPHOTOV.2015.2395146. N. Lenck, A. Dethlefsen, H. Engelmann, K. Nattermann, P. Zachmann, M. Heiss, et al.,

79

[131]

[132]

[133]

[134]

[135]

[136]

Acetic acid production, migration and corrosion effects in ethylene-vinyl-acetate- (EVA-) based PV modules, in: 27th European Photovoltaic Solar Energy Conference and Exhibition, 2012, pp. 2992e2995. https://doi.org/10. 4229/27thEUPVSEC2012-4CO.9.4, in: http:// www.eupvsec-proceedings.com/proceedings? paper¼17673. N.C. Park, W.W. Oh, D.H. Kim, Effect of temperature and humidity on the degradation rate of multicrystalline silicon photovoltaic module, International Journal of Photoenergy (2013). https://doi.org/10.1155/2013/925280. https:// www.hindawi.com/journals/ijp/2013/925280/. E.E. van Dyk, J.B. Chamel, A.R. Gxasheka, Investigation of delamination in an edgedefined film-fed growth photovoltaic module, Solar Energy Materials and Solar Cells 88 (4) (2005) 403e411. https://doi.org/10.1016/j. solmat.2004.12.004. http://www.sciencedirect .com/science/article/pii/S0927024804004866. S.V. Janakeeraman, J. Singh, J. Kuitche, J.K. Mallineni, G. TamizhMani, A statistical analysis on the cell parameters responsible for power degradation of fielded pv modules in a hot-dry climate, in: 2014 IEEE 40th Photovoltaic Specialist Conference, PVSC, 2014, pp. 3234e3238. https://doi.org/10.1109/PVSC. 2014.6925624. G. Oreski, Accelerated Indoor Durability Testing of Polymeric Photovoltaic Encapsulation Materials, 2010, p. 77730D. https://doi. org/10.1117/12.860390, in: http://proceedin gs.spiedigitallibrary.org/proceeding.aspx?doi¼ 10.1117/12.860390. M. Chang, H. Chen, C. Chen, C.H. Hsueh, W.J. Hsieh, Water cooking for backsheet and PV module endurance evaluation, in: 2014 IEEE 40th Photovoltaic Specialist Conference, PVSC, 2014, pp. 1997e2000. https://doi.org/ 10.1109/PVSC.2014.6925317. M. Chang, H. Chen, C. Chen, C. Hsueh, W. Hsieh, Evaluation of Backsheet Endurance by [UV-Water Cooking Test] and Module Reliability by [TC-HF-UV-DH] Sequential Test. Tech. Rep., WIP, 2015. https://doi.org/10. 4229/EUPVSEC20152015-5CV.2.7, in: http:// www.eupvsec-proceedings.com/proceedings? paper¼31791.

80

D URABILITY AND R ELIABILITY

OF

P OLYMERS

[137] C.C. Lin, P. J Krommenhoek, S. Watson, X. Gu, Depth profiling of degradation of multilayer photovoltaic backsheets after accelerated laboratory weathering: cross-sectional Raman imaging, Solar Energy Materials and Solar Cells 144 (2016) 289e299. https://doi.org/10. 1016/j.solmat.2015.09.021. [138] X. Gu, P.J. Krommenhoek, C.C. Lin, L.C. Yu, T. Nguyen, S.S. Watson, Depth Profiling of Mechanical Degradation of PV Backsheets after UV Exposure, vol. 9563, International Society for Optics and Photonics, 2015, p. 956305. https://doi.org/10.1117/12.2187171, in: https:// www.spiedigitallibrary.org/conference-proceed ings-of-spie/9563/956305/Depth-profiling-ofmechanical-degradation-of-PV-backsheets-afterUV/10.1117/12.2187171.short. [139] X. Gu, C.C. Lin, P.J. Krommenhoek, Y. Lyu, J.H. Kim, L.C. Yu, et al., Depth Profiling of Chemical and Mechanical Degradation of UV-Exposed PV Backsheets, IEEE, 2016, pp. 0115e0120. https://doi.org/10.1109/PVSC. 2016.7749560. ISBN 978-1-5090-2724-8, http://ieeexplore.ieee.org/document/7749560/. [140] G. Oreski, G.M. Wallner, Aging mechanisms of polymeric films for PV encapsulation, Solar Energy 79 (6) (2005) 612e617. https://doi.org/ 10.1016/j.solener.2005.02.008. http://www. sciencedirect.com/science/article/pii/S003809 2X05000794. [141] K. Kanuga, Degradation of polyester film exposed to accelerated indoor damp heat aging, in: Photovoltaic Specialists Conference (PVSC), 2011 37th IEEE, IEEE, 2011, pp. 000096e000100. http://ieeexplore.ieee. org/xpls/abs_all.jsp?arnumber¼6185853. [142] N.H. Phillips, B.L. Givot, W. O’Brien, Understanding Polymer Degradation under Voltage Stress in PV Applications, IEEE, 2013, pp. 1585e1588. https://doi.org/10.1109/ PVSC.2013.6744448. ISBN 978-1-47993299-3 978-1-4799-3298-6, http://ieeexplore. ieee.org/document/6744448/. [143] W.J. Gambogi, Y. Heta, J.G. Kopchick, T. Felder, S.W. MacMaster, A.Z. Bradley, et al., Assessment of PV module durability using accelerated and outdoor performance analysis and comparisons, in: 2014 IEEE 40th Photovoltaic Specialist Conference, PVSC, 2014,

AND

OTHER M ATERIALS

[144]

[145]

[146]

[147]

[148]

[149]

IN

P HOTOVOLTAIC M ODULES

pp. 2176e2181. https://doi.org/10.1109/ PVSC.2014.6925356. N. Kim, H. Kang, K.J. Hwang, C. Han, W.S. Hong, D. Kim, et al., Study on the degradation of different types of backsheets used in PV module under accelerated conditions, Solar Energy Materials and Solar Cells 120 (2014) 543e548. https://doi.org/10.1016/j. solmat.2013.09.036. http://linkinghub.elsevier. com/retrieve/pii/S092702481300514X. A. Iguchi, H. Hirabayashi, K. Yamada, H. Nishimura, Evaluation of accelerated-aged PET film for PV backsheets from mechanical strength and partial discharge, Kagaku Kogaku Ronbunshu 40 (3) (2014) 255e260. https://doi. org/10.1252/kakoronbunshu.40.255. http://jlc. jst.go.jp/DN/JST.JSTAGE/kakoronbunshu/40. 255?lang¼en&from¼CrossRef&type¼abs tract. C.C. Lin, Y. Lyu, D.L. Hunston, J.H. Kim, K.T. Wan, D.L. Stanley, et al., Cracking and Delamination Behaviors of Photovoltaic Backsheet after Accelerated Laboratory Weathering, 2015, p. 956304. https://doi.org/ 10.1117/12.2188557, in: http://proceedings. spiedigitallibrary.org/proceeding.aspx?doi¼ 10.1117/12.2188557. C.C. Lin, Y. Lyu, L.C. Yu, X. Gu, Correlation Between Mechanical and Chemical Degradation after Outdoor and Accelerated Laboratory Aging for Multilayer Photovoltaic Backsheets, vol. 9938, International Society for Optics and Photonics, 2016, p. 99380H. https://doi.org/10. 1117/12.2238216, in: https://www.spiedigital library.org/conference-proceedings-of-spie/99 38/99380H/Correlation-between-mechanicaland-chemical-degradation-after-outdoor-andaccelerated/10.1117/12.2238216.short. G. Illya, V. Handara, L. Yujing, R. Shivakumar, A.S. Budiman, Backsheet degradation under salt damp heat environments e enabling novel and innovative solar photovoltaic systems design for tropical regions and sea close areas, Procedia Engineering 139 (2016) 7e14. https:// doi.org/10.1016/j.proeng.2015.09.233. http:// linkinghub.elsevier.com/retrieve/pii/S1877705 815033111. A.A. Lefebvre, G.S. O’Brien, B.S. Douglas, J.D. Knapp, D. Garcia, G. Moeller, Durability

3: D EGRADATION S CIENCE

[150]

[151]

[152]

[153]

[154]

[155]

AND

PATHWAYS

IN

PV S YSTEMS

of Photovoltaic Backsheet Outer Weatherable Layers, IEEE, 2015, pp. 1e5. https://doi.org/ 10.1109/PVSC.2015.7355704. ISBN 978-14799-7944-8, http://ieeexplore.ieee.org/docu ment/7355704/. S. Ponce-Alcantara, A.V. Arang, G.S. Plaza, The importance of optical characterization of PV backsheets in improving solar module power, in: Photovoltaics International, vol. 26, 2015. https://www.pv-tech.org/technical-pape rs/the-importance-of-optical-characterizationof-pv-backsheets-in-improving-so. D.A. Gordon, A. Gok, C.W. Meyer, C.L. Fagerholm, N.W. Sweet, L. Denoyer, L.S. Bruckman, R.H. French. Characterizing the weathering induced haze formation and gloss loss of Poly(ethylene-terephthalate) via MaPd: RTS Spectroscopy, Proc. SPIE 9938, Reliability of Photovoltaic Cells, Modules, Components, and Systems IX, San Diego, CA. August 2016, https://doi.org/10.1117/12.2238295. A. Fairbrother, S. Julien, K.T. Wan, L. Ji, K. Boyce, S. Merzlic, et al., Degradation Analysis of Field-Exposed Photovoltaic Modules with Non-fluoropolymer-based Backsheets, vol. 10370, International Society for Optics and Photonics, 2017, p. 1037003. https://doi.org/10.1117/12.2272488, in: https:// www.spiedigitallibrary.org/conference-procee dings-of-spie/10370/1037003/Degradation-an alysis-of-field-exposed-photovoltaic-moduleswith-non-fluoropolymer/10.1117/12.2272488. short. B.A. Sultan, E. So¨rvik, Thermal degradation of EVA and EBAA comparison. I. Volatile decomposition products, Journal of Applied Polymer Science 43 (9) (1991) 1737e1745. https://doi.org/10.1002/app.1991.070430917. http://onlinelibrary.wiley.com/doi/10.1002/app. 1991.070430917/abstract. F.J. Pern, Luminescence and absorption characterization of ethylene-vinyl acetate encapsulant for PV modules before and after weathering degradation, 00036, Polymer Degradation and Stability 41 (2) (1993) 125e139. https://doi.org/10.1016/0141-3910 (93)90035-H, http://www.sciencedirect.com/ science/article/pii/014139109390035H. D. Berman, S. Biryukov, D. Faiman, EVA laminate browning after 5 years in a grid-

81

[156]

[157]

[158]

[159]

[160]

[161]

connected, mirror-assisted, photovoltaic system in the Negev desert: effect on module efficiency, Solar Energy Materials and Solar Cells 36 (4) (1995) 421e432. https://doi.org/ 10.1016/0927-0248(94)00198-7. http://www. sciencedirect.com/science/article/pii/0927024 894001987. P. Klemchuk, M. Ezrin, G. Lavigne, W. Holley, J. Galica, S. Agro, Investigation of the degradation and stabilization of EVA-based encapsulant in field-aged solar energy modules, Polymer Degradation and Stability 55 (3) (1997) 347e365. https://doi.org/10.1016/ S0141-3910(96)00162-0. http://www.science direct.com/science/article/pii/S014139109600 1620. M. Rodriguez-Vazquez, C.M. Liauw, N.S. Allen, M. Edge, E. Fontan, Degradation and stabilisation of poly(ethylene-stat-vinyl acetate): 1 Spectroscopic and rheological examination of thermal and thermo-oxidative degradation mechanisms, Polymer Degradation and Stability 91 (1) (2006) 154e164. https://doi.org/10.1016/j.polymdegradstab. 2005.04.034. http://www.sciencedirect.com/ science/article/pii/S0141391005002065. C. Peike, L. Purschke, K.A. Weiss, M. Khl, M. Kempe, Towards the origin of photochemical EVA discoloration, in: 2013 IEEE 39th Photovoltaic Specialists Conference, PVSC, 2013, pp. 1579e1584. https://doi.org/10.1109/ PVSC.2013.6744447. J.H. Wohlgemuth, M.D. Kempe, D.C. Miller, Discoloration of PV encapsulants, in: 2013 IEEE 39th Photovoltaic Specialists Conference, PVSC, 2013, pp. 3260e3265. https://doi. org/10.1109/PVSC.2013.6745147. K.A. Berger, U. Jahn, T. Heimberg, S. Altmann, M. Kntges, Mean degradation rates in PV systems for various kinds of PV module failures, in: 32nd European Photovoltaic Solar Energy Conference and Exhibition, 2016, pp. 1435e1443. https://doi.org/10.4229/EUP VSEC20162016-5DP.1.2, in: http://www.eupv sec-proceedings.com/proceedings?paper¼38 352. F.J. Pern, Factors that affect the EVA encapsulant discoloration rate upon accelerated exposure, Solar Energy Materials and Solar Cells 41 (1996) 587e615. https://doi.org/10.

82

[162]

[163]

[164]

[165]

[166]

[167]

[168]

D URABILITY AND R ELIABILITY

OF

P OLYMERS

1016/0927-0248(95)00128-X. http://www. sciencedirect.com/science/article/pii/0927024 89500128X. S. Agro, J. Galica, W.H. Holley, R.S. Yorgensen, Case histories of EVA encapsulant discoloration in fielded modules, AIP Conference Proceedings 306 (1) (1994) 586e596. https://doi. org/10.1063/1.45731. http://aip.scitation.org/ doi/abs/10.1063/1.45731. M.D. Kempe, Ultraviolet light test and evaluation methods for encapsulants of photovoltaic modules, Solar Energy Materials and Solar Cells 94 (2) (2010) 246e253. https://doi.org/ 10.1016/j.solmat.2009.09.009. http://www. sciencedirect.com/science/article/pii/S092702 4809003274. J.S. Jeong, N. Park, Field Discoloration Analysis and UV/temperature Accelerated Degradation Test of EVA for PV, IEEE, 2013, pp. 3010e3013. https://doi.org/10.1109/PVSC. 2013.6745095. ISBN 978-1-4799-3299-3 9781-4799-3298-6, http://ieeexplore.ieee.org/ document/6745095/. T.C. Felder, W.J. Gambogi, J.G. Kopchick, R.S. Peacock, K.M. Stika, T.J. Trout, et al., Optical properties of PV backsheets: key indicators of module performance and durability, in: Proc. SPIE 9179, Reliability of Photovoltaic Cells, Modules, Components, and Systems VII, vol. 9179, 2014. https://doi.org/10.1117/12. 2062063, 91790Pe91796Pe6, https://doi.org/ 10.1117/12.2062063. F. Liu, L. Jiang, S. Yang, Ultra-violet degradation behavior of polymeric backsheets for photovoltaic modules, Solar Energy 108 (2014) 88e100. https://doi.org/10.1016/j.solener.2014. 06.027. http://linkinghub.elsevier.com/retrie ve/pii/S0038092X14003260. A. Jentsch, K.J. Eichhorn, B. Voit, Influence of typical stabilizers on the aging behavior of EVA foils for photovoltaic applications during artificial UV-weathering, Polymer Testing 44 (Supplement C) (2015) 242e247. https://doi. org/10.1016/j.polymertesting.2015.03.022. http://www.sciencedirect.com/science/article/ pii/S0142941815000768. M. Koehl, A. Ballion, Y.-H. Lee, H.-S. Wu, K. Scott, S. Glick, et al., Ultraviolet Radiation Round-Robin Testing of Various Backsheets for Photovoltaic Modules, IEEE, 2015,

AND

OTHER M ATERIALS

[169]

[170]

[171]

[172]

[173]

[174]

[175]

[176]

IN

P HOTOVOLTAIC M ODULES

pp. 1e4. https://doi.org/10.1109/PVSC.2015. 7355603. ISBN 978-1-4799-7944-8, http:// ieeexplore.ieee.org/document/7355603/. N. Kim, S. Lee, X.G. Zhao, D. Kim, C. Oh, H. Kang, Reflection and durability study of different types of backsheets and their impact on c-Si PV module performance, Solar Energy Materials and Solar Cells 146 (2016) 91e98. https://doi.org/10.1016/j.solmat.2015.11.038. http://linkinghub.elsevier.com/retrieve/pii/S09 27024815006315. A. Bouaichi, A.A. Merrouni, A. El Hassani, Z. Naimi, B. Ikken, A. Ghennioui, et al., Experimental evaluation of the discoloration effect on PV-modules performance drop, Energy Procedia 119 (Supplement C) (2017) 818e827. https://doi.org/10.1016/j.egypro. 2017.07.107. http://www.sciencedirect.com/ science/article/pii/S187661021732653X. E. Molenbroek, D.W. Waddington, K.A. Emery, Hot spot susceptibility and testing of PV modules, in: The Conference Record of the Twenty-Second IEEE Photovoltaic Specialists Conference e 1991, vol. 1, 1991, pp. 547e552. https://doi.org/10.1109/PVSC. 1991.169273. D. DeGraaff, R. Lacerda, Z. Campeau, Degradation Mechanisms in Si Module Technologies Observed in the Field; Their Analysis and Statistics, 2011. M. Sadok, B. Benyoucef, M. Benmedjahed, Assessment of PV modules degradation based on performances and visual inspection in Algerian sahara, International Journal of Renewable Energy Resources 6 (1) (2016) 106e116. J.K. Hastings, M.A. Juds, C.J. Luebke, B. Pahl, A study of ignition time for materials exposed to DC arcing in PV systems, in: 2011 37th IEEE Photovoltaic Specialists Conference, 2011, pp. 003724e003729. https://doi.org/10.1109/ PVSC.2011.6185959. J. Strauch, M. Quintana, J. Granata, W. Bower, S. Kuszmaul, Solar Module Arc Fault Modeling at Sandia National Laboratories, 2011. http://energy.sandia.gov/wp-content/ gallery/uploads/Strauch-SNL-Module_Arc_ Fault_Modeling-SAND2010-5881C.pdf. F. Reil, A. Sepanski, W. Vaaen, B. Heeckeren, S. Raubach, M. Vosen, et al., Determination of

3: D EGRADATION S CIENCE

[177]

[178]

[179]

[180]

[181]

[182]

AND

PATHWAYS

IN

PV S YSTEMS

arcing risks in PV modules with derivation of risk minimization measures, in: 2013 IEEE 39th Photovoltaic Specialists Conference, PVSC, 2013, pp. 3159e3164. https://doi.org/ 10.1109/PVSC.2013.6745126. C. Hagendorf, S. Swatek, M. Werner, S. Richter, Understanding the snail trail effect in silicon solar modules on microstructural scale, in: 27th European Photovoltaic Solar Energy Conference and Exhibition, 2012, pp. 3439e3441. https://doi.org/10.4229/ 27thEUPVSEC2012-4BV.3.7, in: http://www. eupvsec-proceedings.com/proceedings?paper¼ 21086. B. Stegemann, P. Grunow, S. Koch, M. Hanusch, S. Wendlandt, A. Bttcher, et al., Electrochemical corrosion within solar panels, in: 27th European Photovoltaic Solar Energy Conference and Exhibition, 2012, pp. 3511e3517. https://doi.org/10.4229/27th EUPVSEC2012-4BV.3.37, in: http://www. eupvsec-proceedings.com/proceedings?paper ¼16731. S. Meyer, S. Richter, S. Timmel, M. Glser, M. Werner, S. Swatek, et al., Snail trails: root cause analysis and test procedures, Energy Procedia 38 (2013) 498e505. https://doi.org/ 10.1016/j.egypro.2013.07.309. http://linking hub.elsevier.com/retrieve/pii/S187661021301 3957. N. Kim, K.J. Hwang, D. Kim, J.H. Lee, S. Jeong, D.H. Jeong, Analysis and reproduction of snail trails on silver grid lines in crystalline silicon photovoltaic modules, Solar Energy 124 (Supplement C) (2016) 153e162. https://doi.org/10.1016/j.solener.2015.11.040. http://www.sciencedirect.com/science/article/ pii/S0038092X1500657X. J. Fan, D. Ju, X. Yao, Z. Pan, M. Terry, W. Gambogi, et al., Study on snail trail formation in PV module through modeling and accelerated aging tests, in: Solar Energy Materials and Solar Cells, vol. 164 (Supplement C), 2017, pp. 80e86. https://doi.org/10.1016/j. solmat.2017.02.013. http://www.sciencedirect. com/science/article/pii/S0927024817300661. S. Meyer, S. Timmel, S. Richter, M. Werner, M. Glser, S. Swatek, et al., Silver nanoparticles cause snail trails in photovoltaic modules, Solar Energy Materials and Solar Cells 121 (2014)

83

[183]

[184]

[185]

[186]

[187]

[188]

171e175. https://doi.org/10.1016/j.solmat. 2013.11.013. http://www.sciencedirect.com/ science/article/pii/S0927024813005849. S. Meyer, S. Timmel, U. Braun, C. Hagendorf, Polymer foil additives trigger the formation of snail trails in photovoltaic modules, Energy Procedia 55 (2014) 494e497. https://doi.org/ 10.1016/j.egypro.2014.08.014. http://www. sciencedirect.com/science/article/pii/S187661 0214012387. H.C. Liu, C.T. Huang, W.K. Lee, S.S. Yan, F.M. Lin, A defect formation as snail trails in photovoltaic modules, Energy and Power Engineering 07 (08) (2015) 348e353. https://doi. org/10.4236/epe.2015.78032. http://www. scirp.org/journal/doi.aspx?DOI¼10.4236/epe. 2015.78032. A. Dolara, G.C. Lazaroiu, S. Leva, G. Manzolini, L. Votta, Snail trails and cell microcrack impact on PV module maximum power and energy production, IEEE Journal of Photovoltaics 6 (5) (2016) 1269e1277. https:// doi.org/10.1109/JPHOTOV.2016.2576682. P. Hacke, M. Kempe, K. Terwilliger, S. Glick, N. Call, S. Johnston, et al., Characterization of Multicrystalline Silicon Modules with System Bias Voltage Applied in Damp Heat, 2010. Valencia, Spain, https://www.nrel.gov/docs/ fy11osti/49344.pdf. W. Herrmann, M. Schweiger, G. Mathiak, Potential-induced degradation - comparison of different test methods and low irradiance performance measurements, in: 27th European Photovoltaic Solar Energy Conference and Exhibition, 2012, pp. 3157e3162. https://doi. org/10.4229/27thEUPVSEC2012-4DO.6.3, in: http://www.eupvsec-proceedings.com/procee dings?paper¼18660. N. Shiradkar, E. Schneller, N.G. Dhere, Finite Element Analysis Based Model to Study the Electric Field Distribution and Leakage Current in PV Modules under High Voltage Bias, vol. 8825, International Society for Optics and Photonics, 2013, p. 88250G. https://doi.org/10. 1117/12.2024740, in: https://www.spiedigit allibrary.org/conference-proceedings-of-spie/ 8825/88250G/Finite-element-analysis-basedmodel-to-study-the-electric-field/10.1117/12. 2024740.short.

84

D URABILITY AND R ELIABILITY

OF

P OLYMERS

[189] P. Hacke, R. Smith, K. Terwilliger, G. Perrin, B. Sekulic, S. Kurtz, Development of an IEC test for crystalline silicon modules to qualify their resistance to system voltage stress: IEC test for c-Si modules, Progress in Photovoltaics: Research and Applications 22 (7) (2014) 775e783. https://doi.org/10.1002/pip.2434. http://doi.wiley.com/10.1002/pip.2434. [190] D. Lausch, V. Naumann, A. Graff, A. Hhnel, O. Breitenstein, C. Hagendorf, et al., Sodium outdiffusion from stacking faults as root cause for the recovery process of potential-induced degradation (PID), Energy Procedia 55 (2014) 486e493. https://doi.org/10.1016/j.egypro. 2014.08.013. http://www.sciencedirect.com/ science/article/pii/S1876610214012375. [191] W. Luo, Y. Sheng Khoo, P. Hacke, V. Naumann, D. Lausch, P. Harvey S, et al., Potential-induced degradation in photovoltaic modules: a critical review, Energy & Environmental Science 10 (1) (2017) 43e68. https://doi.org/10.1039/ C6EE02271E. http://pubs.rsc.org/en/Content/ ArticleLanding/2017/EE/C6EE02271E. [192] G.R. Mon, R.G. Ross Jr., G. Whitla, J. Orehotsky, Predicting Electrochemical Breakdown in Terrestrial Photovoltaic Modules, 1984, pp. 682e692. http://adsabs.harvard. edu/abs/1984pvsp.conf..682M. [193] M. Zentgraf, D. Koshnicharov, T. Roth, R. Lw, W.M. Gnehr, J. Hattendorf, Potential-induced degradation and temperature-driven regeneration: a realistic simulation, in: 28th European Photovoltaic Solar Energy Conference and Exhibition, 2013, pp. 3303e3308. https://doi. org/10.4229/28thEUPVSEC2013-4AV.5.38, in: http://www.eupvsec-proceedings.com/proc eedings?paper¼21471. [194] J. Schneider, C. Ehrich, M. Schak, R. Meitzner, A. Apel, S.H. Schulze, Influence of polymer properties on potential induced degradation of PV-modules, in: 28th European Photovoltaic Solar Energy Conference and Exhibition, 2013, pp. 503e507. https://doi.org/10.4229/28thEU PVSEC2013-1CV.2.21, in: http://www. eupvsec-proceedings.com/proceedings? paper¼25044. [195] S. Hoffmann, M. Koehl, Effect of humidity and temperature on the potential-induced degradation: effect of humidity and temperature on the PID, Progress in Photovoltaics: Research and

AND

OTHER M ATERIALS

[196]

[197]

[198]

[199]

[200]

[201]

IN

P HOTOVOLTAIC M ODULES

Applications 22 (2) (2014) 173e179. https:// doi.org/10.1002/pip.2238. http://doi.wiley. com/10.1002/pip.2238. M.B. Kontopp, M. Krber, M. Schtze, C. Taubitz, Potential induced degradation: model calculations and correlation between laboratory tests and outdoor occurrence, in: 29th European Photovoltaic Solar Energy Conference and Exhibition, 2014, pp. 2490e 2494. https://doi.org/10.4229/EUPVSEC20 142014-5DO.11.3, in: http://www.eupvsecproceedings.com/proceedings?paper¼28866. T. Kajisa, H. Miyauchi, K. Mizuhara, K. Hayashi, T. Tokimitsu, M. Inoue, et al., Novel lighter weight crystalline silicon photovoltaic module using acrylic-film as a cover sheet, Japanese Journal of Applied Physics 53 (9) (2014) 092302. https://doi.org/10.7567/ JJAP.53.092302. http://iopscience.iop.org/ article/10.7567/JJAP.53.092302/meta. V. Naumann, T. Geppert, S. Groer, D. Wichmann, H.J. Krokoszinski, M. Werner, et al., Potential-induced degradation at interdigitated back contact solar cells, Energy Procedia 55 (2014) 498e503. https://doi.org/10. 1016/j.egypro.2014.08.015. http://www.scie ncedirect.com/science/article/pii/S187661021 4012399. M.D. Kempe, D.C. Miller, D.L. Nobles, K. Sakurai, J. Tucker, J.G. Bokria, et al., Development of a Resistivity Standard for Polymeric Materials Used in Photovoltaic Modules, 2015, p. 956302. https://doi.org/10. 1117/12.2189662, in: http://proceedings.spie digitallibrary.org/proceeding.aspx?doi¼10.11 17/12.2189662. A. Halm, A. Schneider, V.D. Mihailetchi, L.J. Koduvelikulathu, L.M. Popescu, G. Galbiati, et al., Potential-induced degradation for encapsulated n-type IBC solar cells with front floating emitter, Energy Procedia 77 (2015) 356e363. https://doi.org/10.1016/j. egypro.2015.07.050. http://www.sciencedirect. com/science/article/pii/S1876610215008188. G.H. Kang, H.B. Kim, T.H. Jung, Y.C. Ju, S.W. Ko, H.E. Song, Prediction of the potential induced degradation of photovoltaic modules based on the leakage current flowing through glass laminated with ethylene-vinly acetate, Journal of Solar Energy Engineering 137 (4)

3: D EGRADATION S CIENCE

[202]

[203]

[204]

[205]

[206]

[207]

[208]

AND

PATHWAYS

IN

PV S YSTEMS

(2015) 041001e041001e6. https://doi.org/10. 1115/1.4029933. https://doi.org/10.1115/1.402 9933. J. Berghold, S. Koch, S. Pingel, S. Janke, A. Ukar, P. Grunow, et al., PID: From Material Properties to Outdoor Performance and Quality Control Counter Measures, vol. 9563, International Society for Optics and Photonics, 2015, p. 95630A. https://doi.org/10.1117/12.2188 464, in: https://www.spiedigitallibrary.org/ conference-proceedings-of-spie/9563/95630A /PID–from-material-properties-to-outdoor-per formance-and-quality/10.1117/12.2188464. short. P. Hacke, P. Burton, A. Hendrickson, S. Spataru, S. Glick, K. Terwilliger, Effects of photovoltaic module soiling on glass surface resistance and potential-induced degradation, in: 2015 IEEE 42nd Photovoltaic Specialist Conference, PVSC, 2015, pp. 1e4. https://doi. org/10.1109/PVSC.2015.7355711. P. Hacke, S. Spataru, K. Terwilliger, G. Perrin, S. Glick, S. Kurtz, et al., Accelerated testing and modeling of potential-induced degradation as a function of temperature and relative humidity, IEEE Journal of Photovoltaics 5 (6) (2015) 1549e1553. https://doi.org/10.1109/ JPHOTOV.2015.2466463. F. Reil, N. Bogdanski, G. Mathiak, Potential induced degradation (PID): field monitoring data evaluation and climatic chamber tests with field modules, in: 32nd European Photovoltaic Solar Energy Conference and Exhibition, 2016, pp. 1898e1900. https://doi.org/10.4229/EUPV SEC20162016-5BV.1.30, in: http://www.eupv sec-proceedings.com/proceedings?paper¼38724. P. Hacke, R. Smith, S. Kurtz, D. Jordan, J. Wohlgemuth, Modeling current transfer from PV modules based on meteorological data, in: 2016 IEEE 43rd Photovoltaic Specialists Conference, PVSC, 2016, pp. 1083e1089. https:// doi.org/10.1109/PVSC.2016.7749780. M. Koehl, S. Hoffmann, Impact of rain and soiling on potential induced degradation, Progress in Photovoltaics: Research and Applications 24 (10) (2016) 1304e1309. https:// doi.org/10.1002/pip.2786. http://onlinelibrary. wiley.com/doi/10.1002/pip.2786/abstract. E. Annigoni, M. Jankovec, F. Galliano, H. Li, L.E. Perret-Aebi, M. Topic, et al., Modeling

85

[209]

[210]

[211]

[212]

[213]

[214]

Potential-Induced Degradation (PID) in Crystalline Silicon Solar Cells: From AcceleratedAging Laboratory Testing to Outdoor Prediction, 2016. http://www.eupvsec-planner.com/ presentations/c38295/modelling_potential-ind uced_degradation_pid_in_crystalline_silicon_ solar_cells_from_accelerated-aging_laborato ry_testing_to_outdoor_prediction.htm. M. Schtze, M. Junghnel, M.B. Koentopp, S. Cwikla, S. Friedrich, J.W. Mller, et al., Laboratory study of potential induced degradation of silicon photovoltaic modules, in: 2011 37th IEEE Photovoltaic Specialists Conference, 2011, pp. 000821e000826. https://doi. org/10.1109/PVSC.2011.6186080. C. Jaeger, A. Graichen, W. Schoeppel, G. Stollwerck, Polyolefin backsheet and new encapsulant supress cell degradation in the module, in: 28th European Photovoltaic Solar Energy Conference and Exhibition, 2013, pp. 3318e3320. https://doi.org/10.4229/28thE UPVSEC2013-4AV.5.44, in: http://www.eupvsecproceedings.com/proceedings?paper¼23901. F. Rummens, Long term accelerated weathering tests on coupons to develop new classes of backsheets, in: 31st European Photovoltaic Solar Energy Conference and Exhibition, 2015, pp. 2478e2481. https://doi.org/10.4229/ EUPVSEC20152015-5CV.2.8, in: http://www. eupvsec-proceedings.com/proceedings?paper¼ 32754. E. Ziegler, K. Sundl, M. Plank, H. Muckenhuber, W. Krumlacher, A.J. Steiner, New thermoplastic, non-curing encapsulation material for PV module applications, in: 28th European Photovoltaic Solar Energy Conference and Exhibition, 2013, pp. 2816e2819. https://doi.org/10.4229/28thEUPVSEC20134CO.9.1, in: http://www.eupvsec-proceedings. com/proceedings?paper¼21303. S. Pop, R. Schulze, X. Wang, B. Yuan, J. Kapur, K. Stika, et al., Ionomer-based PiD-resistant encapsulant for PV modules, in: New Materials and Concepts for Solar Cells and Modules, Amsterdam, The Netherlands, 2014, pp. 152e157. https://doi.org/10.4229/EUPVS EC20142014-1BV.6.37. ISBN 3-936338-34-5. J. Kapur, A. Bennett, J. Norwood, B. Hamzavy, I. Kueppenbender, Tailoring ionomer encapsulants as a low cost solution to potential

86

[215]

[216]

[217]

[218]

[219]

[220]

D URABILITY AND R ELIABILITY

OF

P OLYMERS

induced degradation, in: Material Studies, New Concepts, Ultra-high Efficiency and Space Technology, Paris, France, 2013, pp. 476e479. https://doi.org/10.4229/28thEUPVSEC20131CV.2.10. ISBN 3-936338-33-7. M.C. Lopez-Escalante, L.J. Caballero, F. Martn, M. Gabs, A. Cuevas, J.R. RamosBarrado, Polyolefin as PID-resistant encapsulant material in PV modules, Solar Energy Materials and Solar Cells 144 (2016) 691e699. https://doi.org/10.1016/j.solmat.2015.10.009. http://www.sciencedirect.com/science/article/ pii/S0927024815005206. J.D. Lavandera Antolin, M. Cardinali, C. Liciotti, Temperature dependence of encapsulant volumetric resistivity and influence on potential induced degradation of c-Si modules, in: 29th European Photovoltaic Solar Energy Conference and Exhibition, 2014, pp. 3093e3099. https://doi.org/10.4229/EUPV SEC20142014-5BV.3.32, in: http://www.eupv sec-proceedings.com/proceedings?paper¼28332. K. Wangemann, A. Metz, H. Nagel, Crystalline Si solar cells and modules featuring excellent stability against potential-induced degradation, in: 26th European Photovoltaic Solar Energy Conference and Exhibition, 2011, pp. 3107e 3112. https://doi.org/10.4229/26thEUPV SEC2011-4CO.5.6, in: http://www.eupvsecproceedings.com/proceedings?paper¼14115. V. Naumann, K. Ilse, C. Hagendorf, On the Discrepancy between Leakage Currents and Potential-Induced Degradation of Cristalline Silicon Modules. Tech. Rep., WIP, 2013. https://doi.org/10.4229/28thEUPVSEC20134DO.3.2, in: http://www.eupvsec-proceedings. com/proceedings?paper¼26091. K. Hara, S. Jonai, A. Masuda, Crystalline Si photovoltaic modules functionalized by a thin polyethylene film against potential and dampheat-induced degradation, RSC Advances 5 (20) (2015) 15017e15023. https://doi.org/10. 1039/C4RA13360A. http://xlink.rsc.org/?DO I¼C4RA13360A. H.Y. Hsu, H.H. Hsieh, H.Y. Tuan, J.L. Hwang, Oxidized low density polyethylene: a potential cost-effective, stable, and recyclable polymeric encapsulant for photovoltaic modules, Solar Energy Materials and Solar Cells 94 (6) (2010) 955e959. https://doi.org/10.1016/j.solmat.

AND

OTHER M ATERIALS

[221]

[222]

[223]

[224]

[225]

[226]

[227]

IN

P HOTOVOLTAIC M ODULES

2010.01.020. http://www.sciencedirect.com/ science/article/pii/S0927024810000358. J. Berghold, S. Koch, B. Frohmann, P. Hacke, P. Grunow, Properties of encapsulation materials and their relevance for recent field failures, in: 2014 IEEE 40th Photovoltaic Specialist Conference, PVSC, 2014, pp. 1987e1992. https://doi.org/10.1109/PVSC.2014.6925315. I. Fidalgo, R. Merino, B. Perez, Lamination cycle time optimization using new POE encapsulants, in: New Materials and Concepts for Solar Cells and Modules, WIP, Munich, Germany, 2015, pp. 183e186. https://doi.org/ 10.4229/EUPVSEC20162016-1BV.5.41. C.R. Cording, Optimizing Photovoltaic Module Glass Reliability, vol. 7048, International Society for Optics and Photonics, 2008, p. 70480J. https://doi.org/10.1117/12.795120, in: https://www.spiedigitallibrary.org/confe rence-proceedings-of-spie/7048/70480J/Opti mizing-photovoltaic-module-glass-reliability/ 10.1117/12.795120.short. T.D. Fletcher, N. Phillips, Spread of Flame Testing of Glass - Backsheet C-Si Solar Modules: Degradation Progression and Effect of Module Laminate Materials, IEEE, 2015, pp. 1e6. https://doi.org/10.1109/PVSC.2015. 7355992. ISBN 978-1-4799-7944-8, http:// ieeexplore.ieee.org/document/7355992/. J. Wendt, M. Trger, M. Mette, A. Pfennig, B. Jckel, The link between mechanical stress induced by soldering and micro damages in silicon solar cells, in: Proc of 24th EUPVSEC, 2009, pp. 3420e3423. http://www. q-cells.com/uploads/tx_abdownloads/files/30_ THE_LINK_BETWEEN_MECHANICAL_ STRESS_Paper_04.pdf. M. Kntges, I. Kunze, S. Kajari-Schrder, X. Breitenmoser, B. Bjrneklett, The risk of power loss in crystalline silicon based photovoltaic modules due to micro-cracks, Solar Energy Materials and Solar Cells 95 (4) (2011) 1131e1137. https://doi.org/10.1016/j.solmat. 2010.10.034. http://www.sciencedirect.com/ science/article/pii/S0927024810007129. M. Sander, S. Dietrich, M. Pander, S. Schweizer, M. Ebert, J. Bagdahn, Investigations on Crack Development and Crack Growth in Embedded Solar Cells, vol. 8112, International Society for Optics and Photonics,

3: D EGRADATION S CIENCE

[228]

[229]

[230]

[231]

[232]

[233]

[234]

AND

PATHWAYS

IN

PV S YSTEMS

2011, p. 811209. https://doi.org/10.1117/12. 893662, in: https://www.spiedigitallibrary.org/ conference-proceedings-of-spie/8112/811209/ Investigations-on-crack-development-and-cra ck-growth-in-embedded-solar/10.1117/12.893 662.short. J.I. van Mlken, U.A. Yusufolu, A. Safiei, H. Windgassen, R. Khandelwal, T.M. Pletzer, et al., Impact of micro-cracks on the degradation of solar cell performance based on twodiode model parameters, Energy Procedia 27 (Supplement C) (2012) 167e172. https://doi. org/10.1016/j.egypro.2012.07.046. http://www. sciencedirect.com/science/article/pii/S187661 0212012593. I. Berardone, M. Corrado, M. Paggi, A generalized electric model for mono and polycrystalline silicon in the presence of cracks and random defects, Energy Procedia 55 (2014) 22e29. https://doi.org/10.1016/j.egypro.2014. 08.005. http://www.sciencedirect.com/science/ article/pii/S1876610214012296. M. Paggi, I. Berardone, A. Infuso, M. Corrado, Fatigue degradation and electric recovery in Silicon solar cells embedded in photovoltaic modules, Scientific Reports 4 (2014) srep04506. https://doi.org/10.1038/srep04506. https://www.nature.com/articles/srep04506. J. Kasewieter, F. Haase, M. Kntges, Model of cracked solar cell metallization leading to permanent module power loss, IEEE Journal of Photovoltaics 6 (1) (2016) 28e33. https://doi. org/10.1109/JPHOTOV.2015.2487829. N.S. Shiradkar, E. Schneller, N.G. Dhere, V. Gade, Predicting thermal runaway in bypass diodes in photovoltaic modules, in: 2014 IEEE 40th Photovoltaic Specialist Conference, PVSC, 2014, pp. 3585e3588. https://doi.org/ 10.1109/PVSC.2014.6924881. N. Shiradkar, V. Gade, K. Sundaram, Predicting service life of bypass diodes in photovoltaic modules, in: 2015 IEEE 42nd Photovoltaic Specialist Conference, PVSC), 2015, pp. 1e5. https://doi.org/10.1109/PVSC.2015.7355606. M. Gostein, L. Dunn, Light soaking effects on photovoltaic modules: overview and literature review, in: 2011 37th IEEE Photovoltaic Specialists Conference, 2011, pp. 003126e003131. https://doi.org/10.1109/PVSC.2011.6186605.

87

[235] B. Sopori, P. Basnyat, S. Devayajanam, S. Shet, V. Mehta, J. Binns, et al., Understanding lightinduced degradation of c-Si solar cells, in: 2012 38th IEEE Photovoltaic Specialists Conference, 2012, pp. 001115e001120. https://doi. org/10.1109/PVSC.2012.6317798. [236] J. Lindroos, H. Savin, Review of light-induced degradation in crystalline silicon solar cells, Solar Energy Materials and Solar Cells 147 (Supplement C) (2016) 115e126. https://doi. org/10.1016/j.solmat.2015.11.047. http://www. sciencedirect.com/science/article/pii/S092702 4815006406. [237] M. Kempe, A. Dameron, T. Moricone, M. Reese, Evaluation and Modeling of EdgeSeal Materials for Photovoltaic Applications, IEEE, 2010, pp. 000256e000261. https://doi. org/10.1109/PVSC.2010.5614463. ISBN 9781-4244-5890-5, http://ieeexplore.ieee.org/ document/5614463/. [238] J.A. del Cueto, T.J. McMahon, Analysis of leakage currents in photovoltaic modules under high-voltage bias in the field, Progress in Photovoltaics: Research and Applications 10 (1) (2002) 15e28. https://doi.org/10.1002/pip. 401. http://onlinelibrary.wiley.com/doi/10. 1002/pip.401/abstract. [239] T. McMahon, J. del Cueto, S. Glick, G. Jorgensen, M. Kempe, C. Kennedy, et al., Module Design, Materials, and Packaging Research Team: Activities and Capabilities, 2004. Denver, Colorado, http://citeseerx.ist. psu.edu/viewdoc/download?doi¼10.1.1.605. 799&rep¼rep1&type¼pdf. [240] T. Carlsson, J. Halme, P. Lund, P. Konttinen, Moisture sensor at glass/polymer interface for monitoring of photovoltaic module encapsulants, Sensors and Actuators A: Physical 125 (2) (2006) 281e287. https://doi. org/10.1016/j.sna.2005.07.022. http://www. sciencedirect.com/science/article/pii/S09244 24705004784. [241] J.W. Pankow, S.H. Glick, Plasma surface modification of polymer backsheets: origins of future interfacial barrier/backsheet failure, in: 2006 IEEE 4th World Conference on Photovoltaic Energy Conference, vol. 2, 2006, pp. 2250e2253. https://doi.org/10.1109/ WCPEC.2006.279620.

88

D URABILITY AND R ELIABILITY

OF

P OLYMERS

[242] B. Roeder, E.A. Ermilov, D. Philipp, M. Khl, Observation of Polymer Degradation Processes in Photovoltaic Modules via Luminescence Detection, 2008, p. 70480F. https://doi.org/10. 1117/12.793795, in: http://proceedings. spiedigitallibrary.org/proceeding.aspx?doi¼10. 1117/12.793795. [243] K.A. Weiss, J.P. Huelsmann, T. Kaltenbach, D. Philipp, T. Schuhmacher, J. Wirth, et al., Accelerated Degradation Studies of Encapsulation Polymers, 2008, p. 70480E. https://doi. org/10.1117/12.793609, in: http://proceedings. spiedigitallibrary.org/proceeding.aspx?doi¼10. 1117/12.793609. [244] E.C. Quintana, M.A. Quintana, K.D. Rolfe, K.R. Thompson, P. Hacke, Exploring Diagnostic Capabilities for Application to New Photovoltaic Technologies, IEEE, 2009, pp. 002031e002036. https://doi.org/10.1109/ PVSC.2009.5411471. ISBN 978-1-42442949-3, http://ieeexplore.ieee.org/document/ 5411471/. [245] M. Sander, B. Henke, S. Schweizer, M. Ebert, J. Bagdahn, PV Module Defect Detection by Combination of Mechanical and Electrical Analysis Methods, IEEE, 2010, pp. 001765e001769. https://doi.org/10.1109/ PVSC.2010.5615878. ISBN 978-1-42445890-5, http://ieeexplore.ieee.org/document/ 5615878/. [246] M. Sander, B. Henke, H. Schwarz, S. Dietrich, S. Schweizer, M. Ebert, et al., Characterization of PV Modules by Combining Results of Mechanical and Electrical Analysis Methods, 2010, p. 777308. https://doi.org/10.1117/12.860814, in: http://proceedings.spiedigitallibrary.org/ proceeding.aspx?doi¼10.1117/12.860814. [247] U. Eitner, M. Kontges, R. Brendel, Measuring Thermomechanical Displacements of Solar Cells in Laminates Using Digital Image Correlation, IEEE, 2009, pp. 001280e001284. https://doi.org/10.1109/PVSC.2009.5411248. ISBN 978-1-4244-2949-3, http://ieeexplore. ieee.org/document/5411248/. [248] D.C. Miller, M.D. Kempe, S.H. Glick, S.R. Kurtz, Creep in photovoltaic modules: examining the stability of polymeric materials and components, in: 35th IEEE Photovoltaic

AND

OTHER M ATERIALS

[249]

[250]

[251]

[252]

[253]

[254]

IN

P HOTOVOLTAIC M ODULES

Specialists Conference, IEEE, New York, 2010, pp. 262e268. ISBN 978-1-4244-5891-2. WOS:000287579500053. U. Eitner, M. Kntges, R. Brendel, Use of digital image correlation technique to determine thermomechanical deformations in photovoltaic laminates: measurements and accuracy, Solar Energy Materials and Solar Cells 94 (8) (2010) 1346e1351. https://doi.org/10.1016/j.solmat. 2010.03.028. http://linkinghub.elsevier.com/ retrieve/pii/S0927024810001571. G. TamizhMani, Testing the reliability and safety of photovoltaic modules: failure rates and temperature effects, Photovoltaics International 8 (2010). https://www.pv-tech.org/technicalpapers/testing-the-reliability-and-safety-ofphotovoltaic-modules-failure-rates-an. Z. Xia, J.H. Wohlgemuth, D.W. Cunningham, A Semi-empirical Method of Predicting the Lifetime of EVA Encapsulant and Polyester Based Backsheet Materials, 2009, p. 74120B. https://doi.org/10.1117/12.825472, in: http:// proceedings.spiedigitallibrary.org/proceeding. aspx?doi¼10.1117/12.825472. Z. Xia, J.H. Wohlgemuth, D.W. Cunningham, A Lifetime Prediction of PV Encapsulant and Backsheet via Time Temperature Superposition Principle, IEEE, 2009, pp. 000523e000526. https://doi.org/10.1109/PVSC.2009.5411628. ISBN 978-1-4244-2949-3, http://ieeexplore. ieee.org/document/5411628/. B. Roeder, J. Schlothauer, M. Koehl, Fluorescence Imaging for Analysis of the Degradation of PV-Modules, IEEE, 2011, pp. 003606e003608. https://doi.org/10.1109/ PVSC.2011.6185928. ISBN 978-1-42449965-6 978-1-4244-9966-3 978-1-42449964-9, http://ieeexplore.ieee.org/document/ 6185928/. U. Eitner, S. Kajari-Schrder, M. Kntges, H. Altenbach, Thermal stress and strain of solar cells in photovoltaic modules, in: H. Altenbach, V.A. Eremeyev (Eds.), Shell-like Structures, vol. 15, Springer Berlin Heidelberg, Berlin, Heidelberg, 2011, pp. 453e468. https://doi.org/ 10.1007/978-3-642-21855-2_29. ISBN 978-3642-21854-5 978-3-642-21855-2, http://link. springer.com/10.1007/978-3-642-21855-2_29.

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

[255] S. Kurtz, K. Whitfield, G. TamizhMani, M. Koehl, D. Miller, J. Joyce, et al., Evaluation of high-temperature exposure of photovoltaic modules, Progress in Photovoltaics 19 (8) (2011) 954e965. https://doi.org/10.1002/pip. 1103. WOS:000297938600006. [256] G. Oreski, K. Mller, Qualification of Polymeric Components for Use in PV Modules, vol. 8112, International Society for Optics and Photonics, 2011, p. 81120B. https://doi.org/10.1117/12. 893451, in: https://www.spiedigitallibrary.org/ conference-proceedings-of-spie/8112/81120B/ Qualification-of-polymeric-components-foruse-in-PV-modules/10.1117/12.893451.short. [257] C. Ehrich, S.H. Schulze, I. Hinz, M. Pander, M. Ebert, Development of a Novel Technique for Mechanical Characterization of Polymer Encapsulates in Laminated Glass Beams, 2011. [258] S.H. Schulze, M. Pander, K. Naumenko, A. Girchenko, H. Altenbach, Characterization of polymeric interlayers in laminated glass beams for photovoltaic applications, in: H. Altenbach, V.A. Eremeyev (Eds.), Shell-like Structures, vol. 15, Springer Berlin Heidelberg, Berlin, Heidelberg, 2011, pp. 499e506. https:// doi.org/10.1007/978-3-642-21855-2_32. ISBN 978-3-642-21854-5 978-3-642-21855-2, http:// link.springer.com/10.1007/978-3-642-218552_32. [259] M. Koehl, M. Heck, S. Wiesmeier, Modelling of conditions for accelerated lifetime testing of Humidity impact on PV-modules based on monitoring of climatic data, Solar Energy Materials and Solar Cells 99 (2012) 282e291. https://doi.org/10.1016/j.solmat.2011.12.011. http://www.sciencedirect.com/science/article/ pii/S0927024811006921. [260] C. Peike, T. Kaltenbach, K.A. Wei, M. Koehl, Indoor vs. Outdoor Aging: Polymer Degradation in PV Modules Investigated by Raman Spectroscopy, 2012. https://doi.org/10.1117/12. 929828, p. 84720V, http://proceedings. spiedigitallibrary.org/proceeding.aspx?doi¼10. 1117/12.929828. [261] T.C. Felder, W.J. Gambogi, J.G. Kopchick, L. Amspacher, R.S. Peacock, B. Foltz, et al., Development of backsheet tests and measurements to improve correlation of accelerated exposures to fielded modules, in: Proc. SPIE

89

[262]

[263]

[264]

[265]

[266]

[267]

9563, Reliability of Photovoltaic Cells, Modules, Components, and Systems VIII, vol. 9563, 2015, p. 956303. https://doi.org/10.1117/12. 2188627, 9563, https://doi.org/10.1117/12. 2188627. A. Sinha, O. Sastry, R. Gupta, Detection and characterisation of delamination in PV modules by active infrared thermography, Nondestructive Testing and Evaluation 31 (1) (2016) 1e16. https://doi.org/10.1080/10589759.2015. 1034717. http://www.tandfonline.com/doi/ full/10.1080/10589759.2015.1034717. A. Badiee, I.A. Ashcroft, R.D. Wildman, The thermo-mechanical degradation of ethylene vinyl acetate used as a solar panel adhesive and encapsulant, International Journal of Adhesion and Adhesives 68 (Supplement C) (2016) 212e218. https://doi.org/10.1016/j.ijadhadh. 2016.03.008. http://www.sciencedirect.com/ science/article/pii/S0143749616300549. J. Zhu, G. Surier, D. Wu, D. Montiel-Chicharro, T.R. Betts, R. Gottschalg, Adhesion requirements for photovoltaic modules of polymeric encapsulation, in: 2016 IEEE International Reliability Physics Symposium, IRPS, 2016. https://doi.org/10.1109/IRPS. 2016.7574636. PVe4e1ePVe4e5. R. Shivakumar, S.K. Tippabhotla, V.A. Handara, G. Illya, A.A. Tay, F. Novoa, et al., Fracture mechanics and testing of interface adhesion strength in multilayered structures application in advanced solar PV materials and technology, Procedia Engineering 139 (2016) 47e55. https://doi.org/10.1016/j.proeng. 2015.09.232. http://linkinghub.elsevier.com/ retrieve/pii/S187770581503310X. B. Ottersbck, G. Oreski, G. Pinter, Correlation study of damp heat and pressure cooker testing on backsheets, n/aen/a, Journal of Applied Polymer Science 133 (47) (2016). https://doi. org/10.1002/app.44230, http://onlinelibrary. wiley.com/doi/10.1002/app.44230/abstract. R. Meitzner, S.H. Schulze, Method for determination of parameters for moisture simulations in photovoltaic modules and laminated glass, Solar Energy Materials and Solar Cells 144 (2016) 23e28. https://doi.org/10.1016/j.solmat. 2015.08.014. http://www.sciencedirect.com/ science/article/pii/S0927024815004079.

90

D URABILITY AND R ELIABILITY

OF

P OLYMERS

[268] S. Suzuki, T. Tanahash, T. Doi, A. Masuda, Acceleration of degradation by highly accelerated stress test and air-included highly accelerated stress test in crystalline silicon photovoltaic modules, Japanese Journal of Applied Physics 55 (2) (2016) 022302. https:// doi.org/10.7567/JJAP.55.022302. http://iop science.iop.org/article/10.7567/JJAP.55.02230 2/meta. [269] M. Koehl, S. Hoffmann, S. Wiesmeier, Evaluation of damp-heat testing of photovoltaic modules: evaluation of damp-heat testing of photovoltaic modules, Progress in Photovoltaics: Research and Applications 25 (2) (2017) 175e183. https://doi.org/10.1002/pip.2842. http://doi.wiley.com/10.1002/pip.2842. [270] A. Sinha, R. Gupta, Effects of different excitation waveforms on detection and characterisation of delamination in PV modules by active infrared thermography, Nondestructive Testing and Evaluation 32 (4) (2017) 418e434. https://doi.org/10.1080/10589759. 2016.1265961. https://doi.org/10.1080/105897 59.2016.1265961. [271] X. Sun, T.J. Silverman, Z. Zhou, M.R. Khan, P. Bermel, M.A. Alam, Optics-based approach to thermal management of photovoltaics: selective-spectral and radiative cooling, IEEE Journal of Photovoltaics 7 (2) (2017) 566e574. https://doi.org/10.1109/JPHOTOV.2016.26460 62. http://ieeexplore.ieee.org/document/782 8043/. [272] M.H. Saw, Y.S. Khoo, J.P. Singh, Y. Wang, Cell-to-module optical loss/gain analysis for various photovoltaic module materials through systematic characterization, Japanese Journal of Applied Physics 56 (8S2) (2017) 08MD03. https://doi.org/10.7567/JJAP.56. 08MD03. http://iopscience.iop.org/article/10. 7567/JJAP.56.08MD03/meta. [273] J. Tracy, N. Bosco, F. Novoa, R. Dauskardt, Encapsulation and backsheet adhesion metrology for photovoltaic modules: adhesion in photovoltaic modules, Progress in Photovoltaics: Research and Applications 25 (1) (2017) 87e96. https://doi.org/10.1002/pip. 2817. http://doi.wiley.com/10.1002/pip.2817. [274] C.R. Osterwald, T.J. McMahon, History of accelerated and qualification testing of terrestrial photovoltaic modules: a literature review,

AND

OTHER M ATERIALS

[275]

[276]

[277]

[278]

[279]

[280]

[281]

IN

P HOTOVOLTAIC M ODULES

Progress in Photovoltaics: Research and Applications 17 (1) (2009) 11e33. https://doi.org/ 10.1002/pip.861. http://onlinelibrary.wiley. com/doi/10.1002/pip.861/abstract. S. Kurtz, J. Wohlgemuth, P. Hacke, N. Bosco, M. Kempe, R. Smith, et al., Challenge to Move from ’One Size Fits All’ to PV Modules the Customer Needs., National Renewable Energy Lab. (NREL), Golden, CO (United States), 2011. Tech. Rep. NREL/CP-5200-52672, https://www.osti.gov/biblio/1024514. K. Kato, PVRessQ!: A Research Activity on Reliability of PV Systems from an User’s Viewpoint in Japan, vol. 8112, International Society for Optics and Photonics, 2011, p. 81120K. https://doi.org/10.1117/12.8961 35, in: https://www.spiedigitallibrary.org/ conference-proceedings-of-spie/8112/81120K/ PVRessQ–a-research-activity-on-reliability-ofPV-systems/10.1117/12.896135.short. J.M. Kuitche, G. Tamizh-Mani, R. Pan, Failure Modes Effects and Criticality Analysis (FMECA) Approach to the Crystalline Silicon Photovoltaic Module Reliability Assessment, vol. 8112, 2011. https://doi.org/10.1117/12. 894301, 81120Le81129L, https://doi.org/10. 1117/12.894301. S. Kurtz, J. Wohlgemuth, M. Kempe, N. Bosco, P. Hacke, D. Jordan, et al., Photovoltaic Module Qualification Plus Testing, 2013. Technical Report NREL/TP-5200-60950, https://www. nrel.gov/docs/fy14osti/60950.pdf. M.D. Kempe, Evaluation of the Uncertainty in Accelerated Stress Testing, IEEE, 2014, pp. 2170e2175. https://doi.org/10.1109/ PVSC.2014.6925355. ISBN 978-1-47994398-2 978-1-4799-4399-9, http://ieeexplore. ieee.org/document/6925355/. M. Koehl, Accelerated service life testing of photovoltaic modules, in: C.C. White, J. Martin, J.T. Chapin (Eds.), Service Life Prediction of Exterior Plastics, Springer International Publishing, Cham, 2015, pp. 199e213. https://doi.org/10.1007/978-3319-06034-7_13. ISBN 978-3-319-060330 978-3-319-06034-7, http://link.springer. com/10.1007/978-3-319-06034-7_13. International Electrotechnical Commission, IEC 61215 Terrestrial Photovoltaic (PV) Modules - Design Qualification and Type

3: D EGRADATION S CIENCE

[282]

[283]

[284]

[285]

[286]

[287]

[288]

AND

PATHWAYS

IN

PV S YSTEMS

Approval. International Standard, International Electrotechnical Commission, 2016. https:// webstore.iec.ch/publication/24312. P. Hlsmann, K.A. Weiss, Simulation of water ingress into PV-modules: IEC-testing versus outdoor exposure, Solar Energy 115 (Supplement C) (2015) 347e353. https://doi.org/ 10.1016/j.solener.2015.03.007. http://www. sciencedirect.com/science/article/pii/S00380 92X1500122X. M.D. Kempe, J.H. Wohlgemuth, Evaluation of temperature and humidity on PV module component degradation, in: 2013 IEEE 39th Photovoltaic Specialists Conference, PVSC), 2013, pp. 0120e0125. https://doi.org/10.1109/ PVSC.2013.6744112. E.U. Reisner, G. Stollwerck, H. Peerlings, F. Shafiq, Humidity in a solar moduledhorror vision or negligible?, in: 21st European Photovoltaic Solar Energy Conference, Dresden, Germany, 2006, pp. 2058e2060. G.M. Kimball, S. Yang, A. Saproo, Global acceleration factors for damp heat tests of PV modules, in: IEEE Conference on Photovoltaic Specialists, IEEE, 2016, pp. 0101e0105. https://doi.org/10.1109/PVSC.2016.7749557. ISBN 978-1-5090-2724-8. K. Whitfield, A. Salomon, S. Yang, I. Suez, Damp heat versus field reliability for crystalline silicon, in: 2012 38th IEEE Photovoltaic Specialists Conference, 2012, pp. 001864e001870. https://doi.org/10.1109/PVSC.2012.6317957. J.H. Wohlgemuth, M.D. Kempe, Equating damp heat testing with field failures of PV modules, in: 2013 IEEE 39th Photovoltaic Specialists Conference, PVSC, 2013, pp. 0126e0131. https://doi.org/10.1109/PVSC.2013.6744113. W. Gambogi, Y. Heta, K. Hashimoto, J. Kopchick, T. Felder, S. MacMaster, et al., Weathering and durability of PV backsheets and impact on PV module performance, in: Reliability of Photovoltaic Cells, Modules, Components, and Systems VI, vol. 8825, International Society for Optics and Photonics, 2013, p. 88250B. https://doi.org/10.1117/12. 2024491, in: https://www.spiedigitallibrary.org/ conference-proceedings-of-spie/8825/88250B/ Weathering-and-durability-of-PV-backsheetsand-impact-on-PV/10.1117/12.2024491.short.

91

[289] W. Gambogi, J. Kopchick, T. Felder, S. MacMaster, A. Bradley, B. Hamzavy, et al., Backsheet and module durability and performance and comparison of accelerated testing to long term fielded modules, in: Proceedings of EU-PVSEC, WIP, 2013, pp. 2846e2850. https://doi.org/10.4229/28thEUPVSEC20134CO.10.2, in: http://www.eupvsec-proceedings. com/proceedings?paper¼26097. [290] W. Ko¨ppen, The thermal zones of the earth according to the duration of hot, moderate and cold periods and to the impact of heat on the organic world, Meteorologische Zeitschrift 20 (3) (2011) 351e360. https://doi.org/10.1127/ 0941-2948/2011/105. [291] F. Rubel, M. Kottek, Comments on: “the thermal zones of the earth” by Wladimir ko¨ppen (1884), Meteorologische Zeitschrift 20 (3) (2011) 361e365. https://doi.org/10.1127/ 0941-2948/2011/0258. [292] R.H. Hoyle, Handbook of Structural Equation Modeling, first ed., The Guilford Press, New York, 2012. ISBN 978-1-60623-077-0. [293] K.J. Preacher, E.C. Merkle, The problem of model selection uncertainty in structural equation modeling, Psychological Methods 17 (1) (2012) 1. [294] J.J. Faraway, Linear Models with R, first ed., Chapman and Hall/CRC, 2004. ISBN 1-58488425-8. [295] S. Weisberg, Applied Linear Regression, third ed., Wiley, 2005. ISBN 0-471-66379-4. [296] G. James, D. Witten, T. Hastie, R. Tibshirani, An introduction to statistical learning: with applications in R, in: Springer Texts in Statistics, first ed., Springer, New York, 2013. ISBN 978-1-4614-7137-0. [297] R Core Team, R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing, 2019, p. 00061. [298] J. Ma, N. Wheeler, Y. Xu, W. Du, A. Gok, J. Sun, gSEM: Semi-supervised Generalized Structural Equation Modeling, 2016. https:// cran.r-project.org/web/packages/gSEM/index. html. [299] A. Gelman, J. Hill, Data Analysis Using Regression and Multilevel/Hierarchical Models by Andrew Gelman, 2006. https://doi.org/10. 1017/CBO9780511790942. http://core/books/

92

[300]

[301]

[302]

[303]

[304]

[305]

[306]

[307]

[308]

D URABILITY AND R ELIABILITY

OF

P OLYMERS

data-analysis-using-regression-andmultilevelhierarchical-models/ 32A29531C7FD730C3A68951A17C9D983. P.D. Allison, Fixed Effects Regression Models, SAGE Publications, 2009. ISBN 978-1-48338927-1. Google-Books-ID: 3UxaBQAAQBAJ. R.H. Baayen, D.J. Davidson, D.M. Bates, Mixed-effects modeling with crossed random effects for subjects and items, Journal of Memory and Language 59 (4) (2008) 390e412. https://doi.org/10.1016/j.jml.2007. 12.005. http://www.sciencedirect.com/science/ article/pii/S0749596X07001398. J. Fox, Applied Regression Analysis and Generalized Linear Models, third ed., SAGE Publications, Inc, Los Angeles, 2015. ISBN 978-1-4522-0566-3. A. Gelman, Analysis of variancedwhy it is more important than ever, Annals of Statistics 33 (1) (2005) 1e53. https://doi.org/10.1214/ 009053604000001048. P. Geladi, Chemometrics in spectroscopy. part 1. classical chemometrics, Spectrochimica Acta Part B: Atomic Spectroscopy 58 (5) (2003) 767e782. https://doi.org/10.1016/ S0584-8547(03)00037-5. http://www.science direct.com/science/article/pii/S058485470300 0375. C.M. Andersen, R. Bro, Practical aspects of PARAFAC modeling of fluorescence excitation-emission data, Journal of Chemometrics 17 (4) (2003) 200e215. https://doi.org/ 10.1002/cem.790. http://onlinelibrary.wiley. com/doi/10.1002/cem.790/abstract. K.R. Murphy, C.A. Stedmon, D. Graeber, R. Bro, Fluorescence spectroscopy and multiway techniques, PARAFAC. Analytical Methods 5 (23) (2013) 6557e6566. https://doi. org/10.1039/C3AY41160E. http://pubs.rsc. org/en/content/articlelanding/2013/ay/ c3ay41160e. F. Vogt, B. Dable, J. Cramer, K. Booksh, Recent advancements in chemometrics for smart sensors, Analyst 129 (2004) 492e502. https:// doi.org/10.1039/B401202J. https://doi.org/10. 1039/B401202J. E. Wang, H.E. Yang, J. Yen, S. Chi, C. Wang, Failure modes evaluation of PV module via materials degradation approach, Energy Procedia 33 (2013) 256e264. https://doi.org/10.1016/

AND

OTHER M ATERIALS

[309]

[310]

[311]

[312]

[313]

[314]

[315]

[316]

IN

P HOTOVOLTAIC M ODULES

j.egypro.2013.05.066. http://www.sciencedirect. com/science/article/pii/S1876610213000763. H. Zhou, K.Y. Liang, On estimating the change point in generalized linear models, in: Collections, Beachwood, Institute of Mathematical Statistics, Ohio, USA, 2008, pp. 305e320. ISBN 978-0-940600-73-7. V.M. Muggeo, Segmented, An R package to fit regression models with broken-line relationships, R News 8 (1) (2008) 20e25. H.J. Wenger, C. Jennings, J.J. Iannucci, Carrisa Plains PV power plant performance, in: IEEE Conference on Photovoltaic Specialists, vol. 2, 1990, pp. 844e849. https://doi.org/10.1109/ PVSC.1990.111740. H.J. Wenger, J. Schaefer, A. Rosenthal, B. Hammond, L. Schlueter, Decline of the Carrisa Plains PV power plant: the impact of concentrating sunlight on flat plates, in: The Conference Record of the Twenty-Second IEEE Photovoltaic Specialists Conference e 1991, vol. 1, 1991, pp. 586e592. https://doi. org/10.1109/PVSC.1991.169280. D.D. Sumner, C.M. Whitaker, L.E. Schlueter, Carrisa Plains photovoltaic power plant 1984-7 performance, in: Conference Record of the Twentieth IEEE Photovoltaic Specialists Conference, vol. 2, 1988, pp. 1289e1292. https:// doi.org/10.1109/PVSC.1988.105912. A. Rosenthal, C. Lane, Field-test results for the 6 Mw carrizo solar photovoltaic power-plant, Solar Cells 30 (1e4) (1991) 563e571. http:// apps.isiknowledge.com/full_record.do?product ¼WOS&search_mode¼GeneralSearch&qid¼ 4&SID¼3CH51D78eh5llM1hGdB&page¼ 1&doc¼8. J. Schaefer, L. Schlueter, A. Rosenthal, H. Wenger, Electrical degradation of the Carrisa Plains power plant, in: Tenth E.C. Photovoltaic Solar Energy Conference, Springer, Dordrecht, 1991, pp. 1248e1253. https://doi. org/10.1007/978-94-011-3622-8_315. ISBN 978-94-010-5607-6 978-94-011-3622-8, https://link.springer.com/chapter/10.1007/97894-011-3622-8_315. N.R. Wheeler, Lifetime and Degradation Science of Polymeric Encapsulant in Photovoltaic Systems: Investigating the Role of Ethylene Vinyl Acetate in Photovoltaic Module Performance Loss with Semi-gSEM Analytics, Ph.D.

3: D EGRADATION S CIENCE

AND

PATHWAYS

IN

PV S YSTEMS

thesis, Case Western Reserve University, 2017, https://etd.ohiolink.edu/pg_10?0::NO:10:P10_ ACCESSION_NUM:case1473440287427776. [317] C. Ballif, D.M. Hulji, G. Willeke, A. HesslerWyser, Silver thick-film contacts on highly doped n-type silicon emitters: structural and electronic properties of the interface, Applied Physics Letters 82 (12) (2003) 1878e1880. https://doi.org/10.1063/1.1562338. http://aip. scitation.org/doi/abs/10.1063/1.1562338. [318] Z.G. Li, L. Liang, A.S. Ionkin, B.M. Fish, M.E. Lewittes, L.K. Cheng, et al., Microstructural comparison of silicon solar cells front-side Ag contact and the evolution of current conduction mechanisms, Journal of Applied Physics 110

93

(7) (2011) 074304. https://doi.org/10.1063/1. 3642956. http://scitation.aip.org/content/aip/ journal/jap/110/7/10.1063/1.3642956. [319] A. Gok, Degradation Pathway Models of Poly(ethylene-Terephthalate) under Accelerated Weathering Exposures, Ph.D. thesis, Case Western Reserve University, 2016, https://etd. ohiolink.edu/pg_10?213511361713081::NO: 10:P10_ETD_SUBID:110243. [320] G03 Committee, ASTM G154-16, Standard Practice for Operating Fluorescent Ultraviolet (UV) Lamp Apparatus for Exposure of Nonmetallic Materials. Tech. Rep., ASTM International, West Conshohocken, PA, 2016. www.astm.org.

4 Degradation Processes in Photovoltaic Cells Timothy J. Peshek 1, Justin S. Fada 2,3 and Ina T. Martin 4 Photovoltaics and Electrochemical Systems Branch, John H. Glenn Research Center, NASA, Cleveland, Ohio, United States 2 Space and Science Technology Systems Branch, John H. Glenn Research Center, NASA, Cleveland, Ohio, United States 3 Department of Mechanical and Aerospace Engineering, Case Western Reserve University, Cleveland, Ohio, United States 4 The Materials for Opto/Electronics Research and Education (MORE) Center, Case Western Reserve University, Cleveland, Ohio, United States 1

Chapter Points
Degradation to the module power requires an interaction causing cell-level defects.
Degradation of silicon solar cells is dominated by four modes: potential-induced, light– induced, wafer cracking, and metal corrosion.
These modes affect the cells in different ways and may range from almost no loss of power to complete loss of power.

4.1 Introduction to the Physics of Photovoltaic Devices Over 500 GW of photovoltaics (PV) have been installed globally, at typical costs of greater than $1/W, and PV has become a $300,000,000,000 industry [1,2]. Solar cells are produced in the billion piece volumes. The devices have successfully transferred from the research laboratory [3], to niche applications such as space flight [4], to ultimately a commodity electrical device comprising a sizable percentage of the world’s energy portfolio [5]. We will describe here the most typical mechanisms of solar cell degradation. We will begin by providing a brief introduction into the physics of solar cell performance and an explanation of microscopic limiters of that performance. To better elucidate the concepts of degradation and describe how we understand the microscopic phenomena, we will discuss some typical characterization methods to show the degradation processes. We will then discuss at larger scales the dominant degradation mechanisms of cells including potential-induced degradation, cell cracking, cell corrosion, shunting, and radiation damage.

4.1.1 Semiconductor Growth and Doping Fundamentally, solar cells operate by allowing photo-excited charge carriers to move down a potential energy hill and perform work on an external circuit. This process is accomplished by obtaining materials that have little to no charge conduction in the absence of light; in contrast, the absorption of light allows charge carriers to move freely, specifically by exciting electrons into a conduction state within a semiconductor [6]. Conventional silicon-based solar cells are fabricated from thin (w150 micron) wafers that have been sliced from a large boule. This boule is formed by direct crystallization of the melt in either an active “pulling” process to form single crystal wafers or a passive process of controlled cooling to form multicrystalline wafers [6,7]. For both types of silicon wafers, the boule is “doped” with a very small amount of boron [7,8]. Boron is a group III atom with 1 less valence electron than the neighboring Si atoms and as the crystal grows, the boron substitutes for a silicon atom (denoted as BSi). The crystal becomes slightly positively charged, or p-type, that is the atom is an “acceptor” of an electron. In a similar way, a material can be made n-type by the addition of atoms that have excess electrons compared to the surrounding atoms of the crystal, thereby acting as electron “donors” [7]. Silicon has modest light absorption properties that stem from the nature of its band gap. The difference in energy between the valence band maximum (VBM), which is the highest occupied state at 0 K, and the conduction band minimum (CBM), which is the lowest available electronic state at 0 K, is known as the band gap, Eg, and defines the lowest energy photons that can be absorbed. In silicon, the VBM and CBM are not aligned in the momentum space

Durability and Reliability of Polymers and Other Materials in Photovoltaic Modules. https://doi.org/10.1016/B978-0-12-811545-9.00004-5 Copyright © 2019 Justin S. Fada & Ina T. Martin. Published by Elsevier Inc. All rights reserved. Contribution by Timothy J. Peshek is in public domain.

97

98

D URABILITY AND R ELIABILITY

OF

P OLYMERS

(or ‘k-space’) leading to “indirect” absorption of photons, because the absorption of energies near the band edge also requires a crystal momentum for the electron to fill levels near the CBM. This process is fundamentally less efficient than a direct gap semiconductor that does not require any momentum operation to reach the minimum Eg. Other semiconductors such as gallium arsenide, germanium, or cadmium telluride are direct gap semiconductors and have higher absorption coefficients. However, due to the reversibility of this absorption process, direct gap semiconductors tend also to be more susceptible to carrier recombination and thus more sensitive to crystalline defects that act as scattering centers. The minimum band gap energy, Eg, of the material allows only indirect transitions and requires a photon of energy Eg as well as a phonon to provide a crystalline momentum in order to carry an electron to available conduction states at Eg. Note that the conduction electron has a weak electrostatic attraction to the positive valence hole, and that the motion of these two carriers is bound, and known as a pseudoparticle also called an “exciton”.

4.1.2 pn-Junctions and Photovoltaics The generation of carriers into the conduction band is insufficient to perform electrical work on an external circuit. Without being acted upon by an electric field, the net current through a photoconductive material is zero. In order to perform work, a solar cell must contain a built-in electric field to drive the carriers and provide a voltage drop, which can be accomplished in a semiconductor by fabricating a pn-junction, where p-type and n-type materials interface. At this interface, the excess electrons will mildly diffuse into the region that is deficient in electrons to “compensate” the local charge imbalance, creating a region depleted of excess charge, known, coincidentally, as the “depletion region” [8]. This process cannot continue ad infinitum because the charges are bound to immobile ions held in the crystalline lattice. It is useful to introduce the concept of the Fermi level, which is the energy at which there is a 50% probability of locating an electron. In metals, there is

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

a near continuum of states overlapping valence and conduction levels and the probability of finding an electron in any given level is given by the Fermie Dirac statistics. f ðEÞ ¼

1 1 þ expðE  EF =kTÞ

(4.1)

In a semiconductor, the picture changes somewhat because the valence and conduction bands are separated by an energy gap. The probability of finding a conduction electron then becomes n ¼ Nc f ðEc Þ ¼ Nc 1þexpðEc Ef =kTÞ, and correspondingly the probability of finding a hole in the valence band is p ¼ Nv(1f(Ev)). Therefore for an intrinsic semiconductor, where the donor and acceptor dopants are in approximately equal proportion, the Fermi level is in the middle of the energy gap. When doped, the Fermi level moves towards the valence band for the p-type material, and the conduction band for the n-type material. If the Fermi level crosses into the bands, the semiconductor is said to degenerate and is denoted with a superscript þ or , as in a pþ type. At any junction, the bands will realign so that the Fermi level is pinned across the junction and exists at the same energy with respect to the ionization potential of the material. Thus, in energy space the conduction band of the p-type region appears at a lower energy than the conduction band of the n-type region. This band realignment provides the potential energy for the carriers to perform work because photoexcited electrons from the n-type region will roll downhill into the p-type region and utilize that energy. The electrons are minority carriers in a p-type region; therefore, a solar cell is referred to as a “minority carrier device” [6]. The typical band offsets and alignments for a pn-junction are illustrated in Fig. 4.1.

4.2 Characterization Techniques The properties of the semiconductor and the performance of the devices can be determined by a number of ways. Here we will focus on select relevant characterization techniques that are widely used in understanding solar cell degradation.

4: D EGRADATION P ROCESSES

IN

P HOTOVOLTAIC C ELLS

99

ht

Fermi energy

metal valence band (half full)

lig

metal valence band (half full)

conduction band (empty) negative charge (electron)

forbidden energy band (gap)

positive charge (hole) valence band (full)

metal contact

n-doped semicond.

p-doped semicond.

metal contact

Figure 4.1 Pictorial representation of the band structure and light absorption processes inside an inorganic pn-junction solar cell. Taken from Wikipedia Contributors, LaTeX e Wikipedia, The Free Encyclopedia, 2011. Available from: https://en.wikipedia. org/wiki/Theory_of_solar_cells.

Using Kirchhoff’s laws, we can find that I¼ILeIDIsh, where ID is given by Shockley’s equation and Ish is given by Ohm’s law. From simple substitutions we obtain:
IðVÞ ¼ IL  Id Figure 4.2 The typical “single diode” equivalent circuit of a solar cell. Taken from Wikipedia Contributors, LaTeX e Wikipedia, The Free Encyclopedia, 2011. Available from: https://en. wikipedia.org/wiki/Theory_of_solar_cells.

4.2.1 Current versus Voltage Characteristics The current versus voltage or IeV response is a useful measure of electronic components. For example, a resistor is ohmic, and has a linear IeV curve, whereas a rectifier is a device with a nonlinear IeV curve such as a diode. The presence of the pnjunction in solar cells indicates that they are diodes, commonly called the “diffusion diodes”, and should obey the Shockley equation [10]. For clarity, it is common to illustrate a simple equivalent circuit of a solar cell comprising the discrete electronic components as shown in Fig. 4.2.




  V þ IðVÞRs exp 1 VT

V þ IðVÞRs : Rsh

(4.2)

The ohmic loss mechanisms are contained in resistors: series resistance losses are related to contact resistance, dopant levels in the semiconductor, and resistance of any leads. Parallel resistance losses are parasitic recombination (leakage) currents typically from low resistance points branching the pn-junction, commonly called as “shunts.” The sources of shunts can be structural defects or chemical defects, but tend to be characterized by their energy level that usually exists within the semiconductor band gap. These defect levels are also called “traps” as they tend to be localized spatially and capture charge carriers, which then rapidly recombine and are lost. The losses in a solar cell can be generically referred to as carrier recombination, which can occur by defects or traps (known as the Shockley-Read-Hall mechanism or SRH), Auger recombination, or radiative recombination, where electrons and holes recombine to emit

D URABILITY AND R ELIABILITY

100

OF

P OLYMERS

a characteristic photon [11]. The latter two are uncommon in a typical solar cell operation, thus the SRH recombination tends to dominate efficiency, and is modeled by a parallel “recombination” diode in the circuit as this recombination is nonlinear in voltage. The recombination diode also obeys the Shockley equation, say for a factor of two in the exponential. The equivalent circuit model can still be simplified to include only one diode, where the exponential prefactor for the diffusion and recombination diode are combined into one value, known as the reverse saturation current, I0, and the exponential factors are lumped to create the unfortunately named “ideality factor,” which varies between 1 and 2 typically. For an ideality factor, n ¼ 1 is the “ideal” diode where little recombination occurs in the space-charge region, i.e., the device is not limited by SRH; n ¼ 2 indicates that the recombination in the space charge region dominates performance [6].
IðVÞ ¼ IL  I0 


  V þ IðVÞRs exp 1 nVT

V þ IðVÞRs : Rsh

(4.3)

4.2.2 Imaging and Spectral Techniques Imaging of PV devices has proven to be wellsuited for qualitatively and quantitatively characterizing 2-D spatial properties of cell wafers in a nondestructive manner [12,13]. The primary modes for PV imaging are electroluminescence (EL), photoluminescence (PL), and thermography. These methods involve injecting electrons into the conduction band, which then thermalize to the conduction band minimum energy, and recombine with a hole via a radiative process that emits a characteristic photon equal in energy to the optical band edge. EL imaging utilizes an applied forward-bias to stimulate recombination and photonic emission captured by a camera [14]. Many commercial cameras exist for this purpose with recent academic writings demonstrating the utility of low cost modifications of consumer cameras for PV [15e18]. This instrument is useful for defect detection, shunt detection, comparing metalization parameters, resolving cell level electrical characteristics including series resistance and voltage distributions, discoloration, and cracking [16,17,19e31].

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

PL uses a light source of energy above the band gap incident upon the wafer surface to stimulate carriers. The absorbed light is reemitted at a lower frequency, which is captured by a camera sensor spatially resolving the fluorescence response of the wafer [32e35]. This technique has proven useful for detecting and calculating shunt values, determining carrier lifetimes and net dopant concentrations, determining emitter sheet resistance, tracking EVA aging, among other research questions [20,36e40]. Additionally, it is now possible to perform PL imaging in the field for reliability and lifetime degradation assessment [30]. PL can also be applied in a pulsed form, and the resulting time-dependent luminescence closely monitored. This time-resolved PL is very useful as it is a direct measure of the minority carrier lifetime, s, which determines on average how long photoexcited carriers will exist in the conduction band. Cells with high defect or trap densities will have short minority carrier lifetimes. Thermography is used to detect failure modes such as hot spots and cracks [12]. Thermography records long-wave infrared (LWIR) radiation from local temperature distributions in the field of view. Pulsed thermography can be used to image particular module layers for an insight into the rear-side interconnects, encapsulant bubbles, and front-side interconnects and soldering [12,41]. Utilizing thermographic techniques, such as dark lock-in thermography coupled with EL, has proven useful for deriving local electrical properties of recombination current and series resistance [22,25,42]. Additionally for nondestructive evaluation of encapsulant discoloration, electroluminescent imaging can provide good insights, especially when paired with dark lock-in thermography [43]. This technique provides spatially resolved information on the electrical mismatch due to discoloration showing that mismatch played a key role in causing and accelerating encapsulant discoloration [43]. PL is inherently a spectral technique as the incident wavelength can be varied for more information including the presence of any defect luminescent bands, or calculation of electronehole (exciton) binding energies. In a similar way, absorption spectroscopy measures the percentage of light absorbed by the cell as a function of wavelength, and the external quantum efficiency (EQE) is a measure of the wavelength-dependent short circuit current. In EQE, the integral over all wavelengths must equal the

4: D EGRADATION P ROCESSES

IN

P HOTOVOLTAIC C ELLS

broadband short circuit current obtained from a typical IeV curve. EQE is highly useful because the strength of absorption of light of a specific wavelength also determines the average depth in the wafer where it is absorbed; thus, the EQE can provide depth profiling information, where the short wavelength region is more sensitive to the front surface-related phenomena such as passivation or emitter carrier lifetimes, and the long wavelength is more sensitive to the bulk or rear-surface phenomena.

4.3 Potential-Induced Degradation In the intended application, it is common for PV modules to be strung together to increase the system voltage and reduce copper losses. In this way, system voltages of 1000 V have become commonplace and are typical input voltages for large string inverters, with respect to the ground. The inverter needs to be grounded, and typically will ground one rail of the voltage input: if the negative rail is grounded then the peak system voltage is at þ1000 V; if the positive rail is grounded then the peak voltage is 1000 V. It is also possible to center tap the system so that the system voltage rails are þ500 V and 500 V with respect to the ground. Due to the fact that the electric code requires framing and supports to be grounded as well, the highest potential solar cell may be 1000 V above a nearby frame member, and this electric field has been shown to be responsible for a deleterious cell degradation process. It is noteworthy that two forms of voltage-dependent degradation exist in the literature. For the first one, the passivation layer of high efficiency cells is reversibly degraded, leading to an increase in surface recombination velocity. Another form is more commonly related to large losses of fill factor in nþ/p Al back-surface field modules, which are far more prevalent in the marketplace. We will restrict the discussion here to the latter form as it is more relevant, and less well understood.

4.3.1 Description of Phenomena Potential-induced degradation (PID) has been observed to cause serious degradation effects [44,45]. PID results in high leakage paths between the cells and ground, through the encapsulation and front glass [45]. Early research studies indicated that the onset of PID required the combined effects of soda-lime glass frontsheets, EVA encapsulation, and the solar cell’s

101

antireflection coating (ARC), which is a highly typical combination of commercial solar modules’ attributes. Studying real-world modules has shown that the PID is dependent upon external factors including primarily array voltage, temperature, and humidity [46e50]. The mode of failure is a loss of fill factor, related to reduced shunt resistance when an IeV curve is scanned [48,51]. This observation indicates that there is damage occurring to the pn-junction, and that this damage is manifesting as shunting (a parallel recombination path) within the cells [49]. The presence and location of the shunts can be imaged using dark lock-in thermography [52]. The shunts sink more local current than the surroundings causing I2Rsh power loss and the appearance of hot spots. Interestingly, studies of PID have demonstrated that the amount of degradation in performance is partially reversible by removing the voltage or applying the inverse voltage [51,53]. With the observation and mode of degradation determined, a conventional reliability methodology of attempting to reproduce the behavior in the laboratory was defined [54e58]. Led by the National Renewable Energy Laboratory, researchers proposed accelerated aging conditions for standardized testing of PV modules to show the PID effects [55]. Notably, these tests generally proposed applying high voltage to the shorted PV module voltage rails with respect to the ground so that all of the cells under test would float to the high voltage, e.g., 1000 V, while the frame or sometimes a metal foil placed on the front glass was kept at ground. Periodic IeV curves would indicate the response level and the measurement of the leakage current was also captured. Several different data-driven models of this performance were used by various authors, which we will summarize. Additionally, several researchers performed examinations of the modules to determine the root cause mechanisms of degradation, described in more detail in the following section. These efforts dovetail with empirical information [44,59] obtained from experiments collected by characterization methods including EL and electron microscopy. These data lend support to fundamental scientific mechanisms that have been provided in the literature but are divorced from specific power loss models. The models discussed here are semiempirical in nature, in that they specify a functional form for the data a priori, and subsequently fit model parameters

102

D URABILITY AND R ELIABILITY

OF

P OLYMERS

to the data. The models are generally determined from indoor test data and may be compared or validated against outdoor real-world data on PID [60,61]. This comparison almost necessitates the estimation of an “acceleration factor (AF),” and in conventional reliability engineering the AF is determined by an Arrhenius-form relationship and governed by an activation energy as in chemical kinetics [62]. This form was utilized by Hacke et al. to model the data and determine AF and activation energies for PID so as to cross-compare laboratory and real-world aging conditions [60]. Ea Pmp ¼ 1  Aexp kT RHð%ÞB t2 ; 0 Pmp

(4.4)

where Pmp is the maximum power point of the PV module as a function of time, t, relative humidity, RH, and temperature, T. P0mp is the maximum power point prior to any PID. Ea is the proposed activation energy of the PID process, and A and B are semiempirical fit parameters. Unlike the model by Hacke et al., which relies on an Arrhenius relationship multiplied by a factor that is parabolic in time and a power law in relative humidity, much of the effort in modeling PID has shown the response function for the power loss to be sigmoidal in time [63], and dependent upon voltage intensity, temperature, and humidity. The sigmoidal time dependence vastly differs from the exponential form of Hacke et al. in that the observed saturation of the effect and stabilization of power loss at long times are not modeled. The model developed by Hattendorf is the most complex, and utilizes a sigmoidal form in time [63]. In Hattendorf et al.’s model, the power loss is modeled as a function of applied bias (U), temperature (T ), relative humidity (RH), and time (t). t

Pmp ðU; RH; T; tÞ ¼

P0mp

1  PN

1  es1 1þe

 UU0 2 PN ðUÞ ¼ 1 þ e f  aðRHÞ ¼ RH RH 0

1t0 s2

! (4.5)

(4.6)

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

TT0 q

bðTÞ ¼ e

(4.8)

t0 ðT; RHÞ ¼ abt00

(4.9)

s1 ðTÞ ¼ b2 s01

(4.10)

s2 ¼ s02

(4.11)

Thus the model of Hattendorf utilizes six free parameters that must be empirically fit to the data: U0 ; f; q; s01 ; s02 ; t00 , RH0, and T0 are defined scaling parameters equal to 50% RH and 50 C. A semiempirical model suggested by Taubitz et al. describes a slightly different approach than the earlier models in that it seeks to model the change in shunt resistance, as this variable is the largest contributor to PID losses [64]. The Taubitz model is based on indoor controlled tests and outdoor tests on identical single cell mini modules. However, this model utilizes two distinct forms for the degradation and restoration phases of power change, and utilizes a transition regime for saturation. The shunt resistance in the shunting phase is modeled by: Rsh ¼ as ebs T t

(4.12)

for the regeneration phase by: Rsh ¼ cr þ ar ebr ðTÞT t

(4.13)

and for the transition phase by: Rsh ¼ aT ðTÞðt þ bT ðTÞÞ þ cT

(4.14)

The constants as, ar, aT (T), bs, br(T), bT (T), cR, and cT have to be determined for each specific module type by fitting empirical data. The exponential function in Eq. (4.12) implies that this model cannot describe the stabilization of the degradation seen in the Hattendorf experiments.

4.3.2 Mechanistic Investigations (4.7)

Sodium ions originating from the soda-lime glass were shown to play a prominent role in the degradation process [65,66]. Secondary ion mass spectra

4: D EGRADATION P ROCESSES

IN

P HOTOVOLTAIC C ELLS

showed high sodium accumulation within the ARC of the solar cells that demonstrated PID [67]. Indeed further microscopic studies showed the presence of sodium precipitates on the surface of solar cells. However, the presence of PID in sodium-free glass samples was also reported and indicated that the PID was the result of the motion of any ionic species and is not strictly related to sodium migration [44]. Bauer et al. used EBIC measurements correlated with time of flight secondary ion mass spectrometry (TOF-SIMS) chemical analysis to demonstrate that the shunts are related to localized concentrations of sodium in the SiNx layer on top of the solar cell [68]. These measurements were carried out on laboratory samples of 1-cell minimodules undergoing the conventional PID test. The authors proposed a model of charge flow behavior that hypothetically explains the shunting behavior, whereby a localized high concentration of sodium ions in the ARC would act to pull the local charges from the highly doped emitter layer and leave behind localized p-channels in the emitter that connect the p-type base and the top contact. This proposed mechanism appeared plausible based on the data, but lacked a direct observation of the phenomena. A more direct observation of the probable PID degradation mechanism was provided by Naumann et al. [69]. Those authors demonstrated the prevalence for Na to preferentially decorate the deep channels associated with silicon stacking faults that were native to the multicrystalline solar cell crystallization process. Many of these stacking faults were deep enough to channel carriers directly through the emitter region and bypass the pn-junction altogether [70]. The reversibility of PID suggests that the sodium is physisorbed and still mobile when decorating the stacking faults [69]. The outdiffusion of sodium from the stacking faults under inverse bias is the mechanistic explanation of the reversible shunt resistance change. This mechanism is shown pictorially in Fig. 4.3 where the migration of sodium ions through the glass and EVA encapsulant accumulate on the surface of the cell and decorate stacking faults in the Si bulk. Presently there is no quantitative model for the motion and precipitation of sodium towards the cell and subsequent diffusion into stacking faults, which would be used to predict output power loss from this mechanism.

103

Figure 4.3 Pictorial representation of the hypothetical PID mechanism.

4.3.3 Possible Mitigation Schemes Empirical evidence indicates that at least part of the damage caused by PID can be reversed by applying a subsequent bias reversed to that, which caused the damage [69]. As described above, this effect hypothetically desorbs the sodium ions from stacking fault channels and drives them back towards the ARC and the encapsulant along electrical leakage paths. With the source of shunts now removed, the shunt resistance is now restored to prior values, resulting in an improved fill factor. The hypothesis that the bulk resistivity of the EVA is low enough during operation to allow for a relatively large ionic leakage path can be utilized to suggest better encapsulants that might preclude PID [71e74]. Polyolefin encapsulants have orders of magnitude greater bulk resistivity and during the standardized testing described earlier and they show minimal PID effects [75]. Aside from mitigation of PID, one method of PID elimination is to retain system voltages at thresholds well below where PID would become relevant. Studies have indicated that a PID onset voltage is near 200 V [63]. This argument is in direct contrast to the industry push to higher system voltages in order to minimize copper losses. However, the utilization of microinverters on every module, so that the immediate output of the module is upconverted to line-level alternating current and the outputs of all modules are injected in parallel, would maintain module-level potentials no greater than Voc of the module, typically below 40 V. In this topology, every module has a maximum DC voltage well below the PID damage threshold. There are numerous well-documented benefits of module level power electronics, such as improved energy harvest

104

D URABILITY AND R ELIABILITY

OF

P OLYMERS

and robustness to shading and other current mismatches [76,77] as well as some caveats including understanding system thermal performance and power ratings, [78]. The primary barrier to widespread adoption remains a high cost structure regardless of automated assembly and potential, yet unclear, labor, and soft cost barriers.

4.4 Light-Induced Degradation Light-induced degradation (LID) refers to the loss of efficiency in silicon solar cells that is found during excess carrier injection, for e.g., above-gap light absorption [79]. In the presence of light, silicon solar cells would rapidly lose 3%e4% of their output power by a rapid decrease in the carrier diffusion rate in the bulk. The effect was first reported by Fischer and Pschunder in 1973 for Czochralski-grown solar cells [80]. Several researchers speculated on the cause of the LID in the 70s and 80s, but conclusive results were elusive. This process in Czochralski-grown single crystal wafer might take place in as little as 30 min of sun exposure and then saturate. After this initial drop, the output remains stable [81e83].

4.4.1 Description of Phenomena LID has been observed for single crystal (c-Si) and multicrystalline silicon (mc-Si) wafers, although the effect in mc-Si is smaller due to the lower initial bulk diffusion length [84]. The efficiency of a solar cell is primarily related to two mechanisms: the relative strength of the light absorption to create charge carriers and the ability to collect those carriers at the cell boundaries. The latter is largely determined by the mean minority carrier lifetime, s, which is related to the carrier mobility by the Drude theory. The mechanism for efficiency loss then is the scattering of minority carriers (from defects, lattice phonons, etc.) resulting in recombination of the carrier exciton. s is comprised of many contributors, including bulk and surface recombination contributors, contributors from traps (localized energy states within the energy gap), and contributors from other electrical and structural defects and interfaces, all of which are characterized by their own mean scattering time. The contributor scattering times may be discerned from measurements of s by modeling spectral data from time-resolved PL and EQE. In the case of LID, it was found that the bulk scattering time was being reduced, with a

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

commensurate increase in the trap density of the cells [85]. This measurement explained the loss of efficiency by the increase in probability of carrier recombination, and further led to hypothetical mechanisms such as the presence of deep traps caused by a lattice defect-silver atom complex, or a lattice vacancy-gold atom complex; however, no model was able to successfully explain the reversibility of the phenomenon [81]. Interestingly, later researchers found that a reduction in boron led to better LID response [86,87], and that replacing the dopant boron with the dopant gallium resulted in LID-resistance [88].

4.4.2 Mechanistic Investigations At typical carrier injection levels for solar cells, the minority carrier lifetime is determined by a mechanism known conventionally as the “ShockleyRead-Hall” recombination, that is, the trap-assisted carrier recombination. This lifetime measurement comprises two constituents, a surface recombination lifetime and bulk recombination lifetime. The trap density, NT, can be found from transient capacitance measurements, and the minority carrier lifetime can be measured by various means, but most typically by time-dependent photoluminescence. High fidelity tracking of the minority carrier lifetime over a scale of 105 s of light exposure indicates the presence of two LID responses, which can be modeled by an exponential process with a fast time constant and a slow time constant, shown graphically in Fig. 4.4 [89]. The activation energy for these two time constants, Ea, is obtained by degrading Cz-Si as a function of temperature and fitting the data with the Arrhenius form: Ea

RðTÞ ¼ k0 ekB T :

(4.15)

In Cz-Si, the Ea of fast degradation was found to be 0.23  0.02 eV and the Ea of slow degradation was found to be 0.47  0.035 eV [90]. The activation energies show no dependency on the wafer boron or oxygen concentration, but the preexponential k0 is found to increase with increasing boron doping [90]. Note that the exact mechanism of LID is not well understood, yet strong evidence exists that the underlying root cause is related to formation of oxide complexes with the p-type dopant, which are catalyzed by the absorption of light. That is, the

4: D EGRADATION P ROCESSES

IN

P HOTOVOLTAIC C ELLS

105

Figure 4.4 High fidelity measurements of minority carrier lifetime under LID versus time. From T.U. Naerland, Characterization of Light Induced Degradation in Crystalline Silicon, 2013, Available from: https://brage. bibsys.no/xmlui/handle/11249/249425.

complexes are catalyzed by excess carriers and the mechanism of LID would seemingly still arise if the cells were biased electrically so that excess carriers were injected into the junction by a forward bias. As stated above, early researchers focused on trapproducing defect complexes, but complexes with the crystalline lattice appeared unlikely because of the recoverability of the power losses at modest temperatures, such as 200 C. Hypothetically the trap producing complex candidate involves the substitutional boron dopant, and one early candidate invoked a boroneinterstitial iron complex (BiFei). Although this complex was found to have LID qualities, and be recoverable, the annealing behavior was found to be very different from the earlier work of Fischer and Pschunder, and likely the high iron-content swamped the true LID response. Controlled experiments found that NT varied directly with the concentration of boron [B] for B-doped Si wafers [91]. Although some conflicting reports exist, several researchers have found that the efficiency loss from LID is proportional to the concentration of interstitial oxygen [Oi] and the level of doping present in the solar cell. Thus, the interstitial oxygen hypothetically combines with interstitial boron to form the BeO complexes (BiOi). These complexes change the local charge structure and form deep traps [84,92,93].

Progress in the acquisition of controlled data collection shifted the proposed boroneoxygen complex from BiOi to BsO2i [94,95], indicating the dominant cause of LID is the substitutional p-dopant, which could hypothetically cause the two distinct rates of degradation [96,97]. Later work excluded the requirement of oxygen dimers (O2i) [98] and Voronkov and Falster proposed that the fast rate mechanism utilizes latent BsO2i complexes and the slow rate mechanism utilizes latent XiBsO, where X is a fast-diffusing interstitial impurity with þ1 charge, such as precipitated boron. In this model, BiBsO is formed via BiO dissociation and subsequent Bi reaction with BsO. Thus, the concentration of latent BiBsO is proportional to both p0 and ½Oi 2 , although the complex contains only one oxygen atom [98]. Once known, the kinetics of these deleterious reactions can be modeled in order to generate a deterministic model of performance loss [99,100].

4.4.3 Possible Mitigation Schemes One significant LID mitigation scheme is to eliminate cells based upon p-type bases and switch to an n-type wafer as the base. In this scenario, the base is doped n-type with phosphorus instead of the conventional boron, which eliminates the oxide complexes that reduce carrier scattering times. P-type wafers are the dominant form of solar cell

106

D URABILITY AND R ELIABILITY

OF

P OLYMERS

wafers for primarily historical reasons related to the development of solar cells for space-based power. It was found that p-type wafers were less prone to radiation damage than their n-type counterparts; today, however, the majority of space solar cell technology is based upon multijunction IIIeV materials like GaAs. Generally speaking, p-type wafers with a thin nþ emitter would have the greatest potential for high efficiency because of the enhanced minority carrier lifetime of electrons versus holes. In a p-type base, the electrons are the minority carriers, and are significantly lighter than holes and subsequently benefit from increased mobility. This difference in effective mass is related intimately to the mechanism of conduction within the valence band versus the conduction band. However, modern wafers are far thinner compared to wafers of 20 years ago when the mass manufacture of silicon solar cells was nascent, and not limited strictly by the minority carrier lifetime. Indeed, most effort in improving cells for thin wafers is related to light trapping and increased absorption due to the low IR absorption of silicon. Further, since the LID has been shown to account for a 2%e4% power loss in p-type wafers, which is absent in n-type wafers the prospects for higher efficiency are far greater for n-type wafers [101]. Notably, most high efficiency manufacturers of technologically advanced solar cell architectures such as interdigitated back contact (IBC) and heterojunction intrinsic thin films (HIT) utilize n-type wafers. It is possible to mitigate the effects of LID with a high heat anneal to essentially “deactivate” the boroneoxygen complexes and maintain nearly the initial efficiency [84]. Indeed several p-type wafer manufacturers are investigating this route and developing tools to perform this action. However, those tools will need to have a throughput commensurate with the solar cell production time, and this additional process step will of course add cost to the cell processing [102].

4.5 Solar Cell Cracking Related to the brittleness of silicon as well as the relative thinness of the industrial wafering process, silicon solar cells are highly prone to breakage through cracking. The cracking of the cells may or may not be deleterious, dependent upon the size and direction of the crack(s). Because the cracking tends

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

to be a stochastic process, it is difficult to predict the occurrence, let alone the anticipated power drop. Yet power loss from cracked cells can be significant because the entire substring of a module may require bypassing in order to maintain the string current.

4.5.1 Description of Phenomena Photovoltaic cell cracking results from a variety of thermal and mechanical sources which can originate before, during, and after module manufacture and lead to module power loss [31,103]. A primary driver of cell cracking among any load source is wafer thickness, which has been reduced substantially in recent years increasing susceptibility to cracking [16,104,105]. The first manufacturing process performed on solid silicon ingots is cell wafer slicing. The dominant method for slicing is fixed abrasive wire sawing with a diamond grit coated stainless steel wire [106]. This method has been preferred over slurry sawing due to reduced kerf loss, reduced thickness variation, and increased efficiency; however, this method has the potential to cause microcracking, especially due to the ability to cut thin wafers [106,107]. Providing a means to minimize microcrack formation during this process, the silicon crystal’s h100i orientation cracks in the 20 direction are observed as the most vulnerable with the 60 direction being the least vulnerable to cracking [107]. Interconnects are attached to the wafer surface creating thermo-mechanical stresses that can allow microcracks to form [29]. Critical to the formation of internal stress concentrations and resultantly cracking is the disparity of coefficients of thermal expansion (CTE) chosen for the solder, ribbon, and cell [29]. Additionally, other factors such as the thickness and malleability of the string ribbon contribute to crack formation [29]. The solder melting temperature will define the region when the cell/solder/ribbon system is cooling and therefore possible to build internal stresses from disparate CTEs [29]. Device lamination applies heat and pulls vacuum to melt the polymer encapsulant and compress the system together. Thermal and mechanical variability across the module induce stresses such as from pressing, especially concentrated near the module edges, or from differential temperatures [13]. Improper tuning of this process can result in encapsulant defects, which adversely affect the cell. For

4: D EGRADATION P ROCESSES

IN

P HOTOVOLTAIC C ELLS

instance, trapped gases from undercuring and encapsulant stiffening from overcuring impart local stress concentrations that can effect crack initiation and propagation [13,108]. After leaving the factory, transportation and handling impact induced mechanical cracking from vibratory motion, point impacts, or sudden accelerations [104,109,110]. When transportation simulations are performed using various indoor apparatus, most cracks that appear do not impact the initial power production. However, when transportation simulation is paired with cyclic environmental stressors, Pmax was observed to degrade maximally by 2.5% [110]. This is confirmed by analyses showing that every cell in a module can be cracked, but as long as a region of the cell does not disconnect, a minimal reduction in power from nominal is recorded [13,31]. However, when mechanical loading is induced near a harmonic of the module, cracking and significant power loss of nearly 8% are observed [110]. Additionally, simulations of environmental loading conditions such as wind and snow loads show similar results [31,109e111].

4.5.2 Mechanistic Investigations The spatial and orientation distributions of wafer cracking have been investigated using the IEC 61215 10.16 standard static mechanical load test [112]. EL was used to capture the cracked state observed after applying push/pull loading to a maximum of 5400 Pa for 27 modules [112]. Classification of crack orientations showed that 50% of cracks were parallel to the busbars, 20% were at 45 , 15% had cracks in several directions, 14% were dendritic, and 1% oriented perpendicular to the busbars [112]. Of the two dominant groups, the cells with 45 cracks were concentrated near the corners of the module and the cells with parallel cracks were present in all areas except along the shorter edges. For charge carrier transport from the active semiconductor wafer, metallic ribbons typically composed of tin-coated copper are used. Many variants of heat sources are used to solder ribbons to the cell, such as lasers, IR lamps, hot air, soldering irons, or induction coils [29]. In any case, a heat flux is imparted on the cell surface, which due to the differences in coefficients of thermal expansion (CTE), causes the Si and Cu to differentially expand. As the surface cools after soldering, microcracks can

107

initiate. This issue is enhanced by the reduced thickness of wafers below 200 mm [29]. Of the common solder materials, SnePb and SneAg, Pb-bearing alloys would likely be better for reducing microcracks due to their CTE closer to Cu and Si but also because they have a lower melting point and would be inducing stress for less time during the cooling process (with all else constant) [29]. However, due to environmental concerns with Pb, Ag-based solder is commonly used. The crack separation width has been shown to be critical in power loss [30]. If the crack width is 2 mm or less there is no impact on the grid-line resistance [30]. However, if the crack width becomes 7 mm or greater the finger becomes completely separated [30]. When loading sequences designed to mimic module transportation scenarios were applied to laboratory modules, cracking occurs with varied effects on power loss [109e111]. When transportation or wind gusts were simulated at the resonance frequency of the modules, power loss was observed to reach 8% [109e111]. When cycling vibrational loading was applied with an additional environmental stressor, such as thermal cycling, approximately 1.5%e2.5% power loss was observed [110].

4.5.3 Possible Mitigation Schemes Numerous potential cracking mitigation schemes exist since the root of the problem is so widely variable. The use of thinner and more ductile wires when tabbing can reduce initiation of microcracks while the wafer/solder pair cools beneath the melting point of the solder [29]. Additionally, the aforementioned solder alloys that are better matched to the CTE of Si could be invoked in order to reduce the tendency of microcracking during the soldering process, or a combination of alloy composition and ductility could lead to improvements [104]. Proper curing of the EVA can reduce the stiffening that increases the potential for thermal and mechanical stress [13]. Additionally, moisture barriers could be preventative of cracking during freezeethaw cycles in cold environments. In the case of crystalline silicon wafers, an optimization of the ingot orientation and the wafering saw can be utilized to reduce microcracks [106,107]. Although difficult to track and prevent, many cracks are evidently formed during the physical installation process as well as transportation and

108

D URABILITY AND R ELIABILITY

OF

P OLYMERS

mechanical loading after installation [109]. Clear quality and building codes for installation and a trained workforce including site inspectors could minimize the localized cracking from handling and installation but this concept needs to be better understood to inform the authorities having jurisdiction.

4.6 Cell Metalization Corrosion Although the fundamental discussion of pnjunctions earlier in this chapter covered the basics of solar cell operation, without low resistance metal contacts to the cells there would be significant voltage drops that would limit power. Typical cell metalization utilizes colloidal pastes that are applied to the wafers and then fired in a high temperature furnace to activate. The process of metalization is engineered for optimal cost and speed in production and minimal shadowing (on the front surface) and is not optimized for lifetime performance. Indeed, over time metal corrosion is observed in arrays in the field and in accelerated aging studies. The corrosion chemistry is complicated in general, but the replacement of high conductive metal particles suspended in an interfacial glass thick film with less conductive metal compounds results in high series resistance in the IeV curves and subsequent voltage losses under operation.

4.6.1 Description of Phenomena Cell metalization corrosion has long been known as a major factor contributing to overall module performance degradation [113e121]. The main mechanisms causing metalization corrosion are due to galvanic corrosion of metallic elements composing the cell interfacial interconnects, including the alumina paste layer (APL) rear-surface, PbeSn solder, and silver screen printed silver grid on the frontside surface. Metallic corrosion is initiated in the presence of galvanic couples in solution at these interfaces, with lowest oxidation potential couples forming oxides and corroding first. Pathways for corrosion are dominated by the ingress of moisture from the exterior of the module to the cell surfaces. This causes corrosion directly while also causing hydrolysis of the encapsulant layers, which are most typically ethylene vinyl acetate (EVA), causing acetic acid to accumulate at the surface of the cell leading to a more severe corrosive effect [113].

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Moisture ingress varies based on several factors, which most obviously include the materials chosen, however, manufacturing variations such as in the lamination process where heat flux and lay-up quality control could lead to delamination in the field [13]. The effect on performance degradation has been shown to vary with module construction, whether it be a traditional polymereglass or glasseglass, with the module permeability directly related to the variation in performance [113,114,122]. Characterization of various backsheet materials has shown that temperature, film thickness, and layer composition are the key variables in the acetic acid transmission rate (AATR) [117]. The lowest permeation rates were found for laminates containing a polyester (PET) core and single PET layer [117]. The AATR of the PET core was found to be the determining factor in the permeation rate of the entire backsheet attributed to the core thickness, which is much larger than the laminated film materials [117]. Highest AATR values were found to be for cured EVA leading to the conclusion that there is no retention of acetic acid by the EVA layers in the module [117]. Typically, an additional Al layer would have a significant barrier effect as in the case of oxygen and water vapor permeation; however, for acetic acid this is not the case [117]. From REMEDX and FTIR-ATR, the corrosion of Al can be seen causing pinholes and microcracks that allow permeation [117].

4.6.2 Mechanistic Investigations Corrosion of the cell can be broken down into a series of processes, which govern the macroscopic manifestation of power loss from a module. First, migration of moisture from outside the module to the front side of the cell causes galvanic corrosion of the soldered connections dominated primarily by the Sn and Pb reactions. Corrosion tends to form along the interface of the Ag electrode and the SnePb solder [113]. This corrosion in turn would decrease the electrical connection and increase the electrical series resistance. The low oxidation potentials of SnePb lead to a slower corrosion rate during the early stages of moisture ingress, observed during the first 1000e1500 h of accelerated damp-heat exposure [123]. After substantial moisture ingress, hydrolysis of the EVA copolymer would initiate the release of acetic acid and corrosion of the Ag grid lines along with other metalized contacts.

4: D EGRADATION P ROCESSES

IN

P HOTOVOLTAIC C ELLS

As moisture penetrates through the cell packaging, the most common metal elements to generate galvanic couples on the rear side of the cell are Ag, Cu, Sn, Pb, and Al [113]. These elements have oxidation potentials of þ0.799, þ0.337, 0.136, 0.358, and 1.662 V, respectively [113]. The elements with lower potential values corrode first, which in this case is Al, which in traditional cells is concentrated on the rear side in the alumina paste layer (APL). In H2O, the reactions for anode and cathode are given as: Al/Al3þ þ 3e ðAnodeÞ 2H2 O þ 2e /2OH  þ H2 [:ðCathodeÞ However, the more severe APL corrosion mechanism is due to acetic acid [113]. This reaction is typically given by: 3 Al þ 3H þ /Al3þ þ H2 [: 2

(4.16)

Corrosion in this manner can weaken connections and lead to an increase in series resistance [113]. During module aging tests under 25 C, 85% relative humidity (RH), damp-heat (DH), and acetic acid atmosphere, the acetic acid exposed modules experienced much more severe degradation [113,122]. Analysis of the front side surface shows high concentrations of Sn and Pb. The presence of water and acid implies that the SnePb alloy solder dissolved forming oxides. Reactions of these elements in the presence of water are given by: Sn/Sn2þ þ 2e ðSn AnodeÞ Pb/Pb2þ þ 2e ðPb AnodeÞ 2H2 O þ 2e /2OH  þ H2 [:ðCathodeÞ The more severe corrosion in acetic acid is typically given as [113]: Sn þ 2H þ /Sn2þ þ H2 [

(4.17)

Pb þ 2H þ /Pb2þ þ H2 [:

(4.18)

109 Under the influence of 25 C, 85% RH damp-heat testing of traditional glassepolymer and glasseglass modules using the same materials, the traditional architecture resulted in a considerable decrease in Pmax of 1.2% for the first 1500 h and 1.63% after 2000 h [113]. This is compared to a 0.07%e0.16% degradation rate of Pmax for the glasseglass modules. These results were corroborated by additional researchers [122]. The lamination process during manufacturing is critical for minimizing the potential emergence of corrosion [13,117,118]. The preheating stage of the lay-up requires that the temperature and time applied fall within specific ranges to ensure that the proper gel content is achieved and that overcuring or undercuring do not occur. Undercuring of EVA can lead to formation of voids and higher concentrations of unreacted peroxides, which can cause reliability issues in the field [13,108]. Undercuring also leads to a lower EVA elastic modulus, which could enhance moisture ingress due to delamination [13]. Overcuring is less of a concern for metalization corrosion; however, this has been shown to lead to quicker discoloration of the encapsulant due to the formation of photosensitive chromophores [124]. All mechanisms for cell surface corrosion, in effect, resolve to an adverse effect on the module output power through the increase of series resistance inhibiting the flow of charge carriers out of the semiconductor wafer [125].

4.6.3 Possible Mitigation Schemes Moisture ingress mitigation is key to ensuring corrosion reduction. Water either directly allows galvanic corrosion or indirectly initiates hydrolysis of the encapsulating polymers producing acetic acid causing a more severe corrosive effect [13,113,122] A first line of defense is proper choice of materials (laminates, solder, and ribbon). In the manufacturing process, the most important step shown to have a direct relationship with corrosion is the lamination process [13]. The preheating stage of lamination is critical to allow moisture to escape the encapsulant. This is especially important when there are concerns with the storage procedure or the age of the encapsulant, which could adversely affect the moisture concentration [13]. Additionally, optimization and monitoring

110

D URABILITY AND R ELIABILITY

OF

P OLYMERS

of the preheating stage is important to mitigate the trapping of air bubbles in the encapsulant layer. If air bubbles are not allowed the proper time and temperature to escape the lay-up, there is an increased chance of future delamination and varied thermomechanical stress in the module [13,108]. When the laminating membrane squeezes the lay-up to fuse the layers together, it is common to have reduced material thickness at the lay-up edges. This can accelerate moisture ingress and at the least will induce stress in the glass, which could increase the chance of fracture [13]. Having a properly fitted frame for the module could help mitigate this issue [13]. Controlling preheat flux to the lay-up during lamination is directly related to the reliability of the module in field conditions [13,117,118]. During this process the gel content needs to accurately reach the cure point, which is typically in the 70%e80% range. If a properly cured state is not achieved, the permeation rates have been shown to change considerably and lifetime performance would then tend to suffer [13,108,117,118,125].

4.7 Conclusions The cost of photovoltaic electricity is intimately linked to the lifetime power production of the systems. We reviewed the modes and mechanisms of four categories of common PV power degradation. Specifically, we described how these modes affect the cellular and modular performance and what is purported to be causative on a microscopic scale and how these mechanisms are measured and understood by researchers. It is possible to reverse, mitigate, or eliminate all of these degradation mechanisms through better understanding, scientific inquiry, and further development. If achieved, cell and modulelevel performance may reach truly impressive timescales, unheard of for commodity electronics.

References [1] Trends in Photovoltaic Applications. Available from: http://www.iea-pvps.org/fileadmin/dam/ public/report/statistics/2018_iea-pvps_report_ 2018.pdf. [2] V. Henze, Clean Energy Investment Exceeded $300 Billion Once Again in 2018.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Available from: https://about.bnef.com/blog/ clean-energy-investment-exceeded-300-billion2018. [3] The Invention of the Solar Cell, Available from: https://www.popsci.com/article/science/ invention-solar-cell. [4] Vanguard 1, Page Version ID: 798451799, 2017, Available from: https://en.wikipedia.org/w/ index.php?title¼Vanguard_1&oldid¼798451799. [5] Timeline of solar cells, Page Version ID: 800982587, 2017, Available from: https://en. wikipedia.org/w/index.php?title¼Timeline_of_ solar_cells&oldid¼800982587. [6] M.A. Green, Solar Cells: Operating Principles, Technology, and System Applications, Prentice-Hall Inc., Englewood Cliffs, NJ, 1982. Available from: https://www.osti.gov/scitech/ biblio/6051510. [7] S.M. Sze, K.K. Ng, Physics of Semiconductor Devices, John Wiley & sons, 2006. Available from: https://books.google.com/books?hl¼en& lr¼&id¼o4unkmHBHb8C&oi¼fnd&pg¼PR7& dq¼szeþsemiconductorþdevices&ots¼wGq m9Iuk4X&sig¼H9p72InPIOYMJqijz8gcNR WrYvs. [8] D.A. Neaman, Semiconductor Physics and Devices, Irwin, Chicago, 1992. Available from: https://pdfs.semanticscholar.org/bbe0/84bff4b6 12b7e3acfd33ab1ff376f86e7bfc.pdf. [9] Wikipedia contributors, LaTeX d Wikipedia, The Free Encyclopedia, 2011. Available from: https://en.wikipedia.org/wiki/Theory_of_solar_ cells. [10] W. Shockley, The theory of p-n junctions in semiconductors and p-n junction transistors, Bell Labs Technical Journal 28 (3) (1949) 435e489. Available from: http://onlinelibrary. wiley.com/doi/10.1002/j.1538-7305.1949.tb03 645.x/full. [11] C.T. Sah, R.N. Noyce, W. Shockley, Carrier generation and recombination in p-n junctions and p-n junction characteristics, Proceedings of the IRE 45 (9) (1957) 1228e1243. Available from: http://ieeexplore.ieee.org/abstract/ document/4056679/. [12] Review of Failures of Photovoltaic Modules, Available from: http://iea-pvps.org/index.php? id¼275.

4: D EGRADATION P ROCESSES

IN

P HOTOVOLTAIC C ELLS

[13] M. Kntges, G. Oreski, U. Jahn, M. Herz, P. Hacke, K.-A. Weiss, et al., Assessment of photovoltaic module failures in the field, IEAPVPS Task 13 (2017). T13-09-2017. Available from: http://www.iea-pvps.org/index.php?id¼ 435. [14] T. Fuyuki, H. Kondo, T. Yamazaki, Y. Takahashi, Y. Uraoka, Photographic surveying of minority carrier diffusion length in polycrystalline silicon solar cells by electroluminescence, Applied Physics Letters 86 (26) (June 2005) 262108. Available from: http://aip. scitation.org/doi/10.1063/1.1978979. [15] J.S. Fada, N.R. Wheeler, D. Zabiyaka, N. Goel, T.J. Peshek, R.H. French, Democratizing an electroluminescence imaging apparatus and analytics project for widespread data acquisition in photovoltaic materials, Review of Scientific Instruments 87 (8) (August 2016) 085109. Available from: http://scitation.aip.org/content/aip/ journal/rsi/87/8/10.1063/1.4960180. [16] A.M. Gabor, R. Janock, A. Anselmo, H. Field, Solar Panel Design Factors to Reduce the Impact of Cracked Cells and the Tendency for Crack Propagation, Denver, CO USA, 2015. Available from: https://www.nrel.gov/pv/ assets/pdfs/2015_pvmrw_04_gabor.pdf. [17] J.S. Fada, A. Loach, A.J. Curran, J.L. Braid, S. Yang, T.J. Peshek, et al., Correlation of IeV Curve Parameters with Module-Level Electroluminescent Image Data Over 3000 Hours Damp-Heat Exposure. Washington, D.C., 2017. [18] J.S. Fada, M.A. Hossain, J.L. Braid, S. Yang, T.J. Peshek, R.H. French, Electroluminescent Image Processing and Cell Degradation Type Classification via Computer Vision and Statistical Learning Methodologies. Washington, D.C., 2017. [19] R. Burhenne, O. Stern, M. Bucher, M. Lynass, O. Mayer, M. Zettl, et al., Defect detection in photovoltaic modules using electroluminescence imaging, in: 27th European Photovoltaic Solar Energy Conference and Exhibition, October 2012, pp. 3374e3378. Available from: http://www.eupvsec-proceedings.com/ proceedings?paper¼17445. [20] O. Breitenstein, J. Bauer, T. Trupke, R.A. Bardos, On the detection of shunts in silicon solar cells by photo- and

111

[21]

[22]

[23]

[24]

[25]

[26]

[27]

electroluminescence imaging, Progress in Photovoltaics: Research and Applications 16 (4) (June 2008) 325e330. Available from: http://onlinelibrary.wiley.com/doi/10.1002/ pip.803/abstract. K. Drabczyk, G. Kulesza-Matlak, A. Drygaa, M. Szindler, M. Lipiski, Electroluminescence imaging for determining the influence of metallization parameters for solar cell metal contacts, Solar Energy 126 (Supplement C) (March 2016) 14e21. Available from: http:// www.sciencedirect.com/science/article/pii/ S0038092X15007021. D. Hinken, K. Ramspeck, K. Bothe, B. Fischer, R. Brendel, Series resistance imaging of solar cells by voltage dependent electroluminescence, Applied Physics Letters 91 (18) (October 2007) 182104. Available from: http://aip.scitation.org/ doi/abs/10.1063/1.2804562. M. Glatthaar, J. Haunschild, R. Zeidler, M. Demant, J. Greulich, B. Michl, et al., Evaluating luminescence based voltage images of silicon solar cells, Journal of Applied Physics 108 (1) (July 2010) 014501. Available from: http:// aip.scitation.org/doi/abs/10.1063/1.3443438. H. Kampwerth, T. Trupke, J.W. Weber, Y. Augarten, Advanced luminescence based effective series resistance imaging of silicon solar cells, Applied Physics Letters 93 (20) (November 2008) 202102. Available from: http:// aip.scitation.org/doi/full/10.1063/1.2982588. T. Trupke, E. Pink, R.A. Bardos, M.D. Abbott, Spatially resolved series resistance of silicon solar cells obtained from luminescence imaging, Applied Physics Letters 90 (9) (February 2007) 093506. Available from: http://aip. scitation.org/doi/abs/10.1063/1.2709630. A. Sinha, O.S. Sastry, R. Gupta, Nondestructive characterization of encapsulant discoloration effects in crystalline-silicon PV modules, Solar Energy Materials and Solar Cells 155 (Supplement C) (October 2016) 234e242. Available from: http://www.sciencedirect.com/science/ article/pii/S092702481630191X. T. Potthoff, K. Bothe, U. Eitner, D. Hinken, M. Kntges, Detection of the voltage distribution in photovoltaic modules by electroluminescence imaging, Progress in Photovoltaics: Research and Applications 18 (2) (March

112

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

D URABILITY AND R ELIABILITY

OF

P OLYMERS

2010) 100e106. Available from: http:// onlinelibrary.wiley.com/doi/10.1002/pip.941/ abstract. T.M. Pletzer, J.I. van Mlken, S. Riland, O. Breitenstein, J. Knoch, Influence of cracks on the local currentevoltage parameters of silicon solar cells, Progress in Photovoltaics: Research and Applications 23 (4) (April 2015) 428e436. Available from: http://onlinelibrary.wiley.com/ doi/10.1002/pip.2443/abstract. A. Gabor, M. Ralli, S. Montminy, L. Alegria, C. Bordonaro, J. Woods, et al., Soldering induced damage to thin Si solar cells and detection of cracked cells in modules, in: Proceedings of the 21st European Photovoltaic Solar Energy Conference, January 2006. M. Kntges, S. Kajari-Schrder, I. Kunze, Crack statistic for wafer-based silicon solar cell modules in the field measured by UV fluorescence, IEEE Journal of Photovoltaics 3 (1) (January 2013) 95e101. M. Kntges, I. Kunze, S. Kajari-Schrder, X. Breitenmoser, B. Bjrneklett, The risk of power loss in crystalline silicon based photovoltaic modules due to micro-cracks, Solar Energy Materials and Solar Cells 95 (4) (April 2011) 1131e1137. Available from: http://www.sciencedirect.com/science/article/ pii/S0927024810007129. T. Trupke, R.A. Bardos, Photoluminescence: a surprisingly sensitive lifetime technique, in: Conference Record of the Thirty-First IEEE Photovoltaic Specialists Conference, 2005, 2005, pp. 903e906. T. Trupke, R.A. Bardos, M.C. Schubert, W. Warta, Photoluminescence imaging of silicon wafers, Applied Physics Letters 89 (4) (July 2006) 044107. Available from: http://aip. scitation.org/doi/abs/10.1063/1.2234747. T. Trupke, B. Mitchell, J.W. Weber, W. McMillan, R.A. Bardos, R. Kroeze, Photoluminescence imaging for photovoltaic applications, Energy Procedia 15 (Supplement C) (January 2012) 135e146. Available from: http://www.sciencedirect.com/science/article/ pii/S1876610212003529. T. Trupke, Photoluminescence and electroluminescence characterization in silicon photovoltaics, in: A. Reinders, P. Verlinden, W.V. Sark, A. Freundlich (Eds.), Photovoltaic

AND

OTHER M ATERIALS

[36]

[37]

[38]

[39]

[40]

[41]

[42]

IN

P HOTOVOLTAIC M ODULES

Solar Energy, John Wiley & Sons, Ltd, 2016, pp. 322e338. https://doi.org/10.1002/ 9781118927496.ch30. Available from: http:// onlinelibrary.wiley.com. Y. Augarten, T. Trupke, M. Lenio, J. Bauer, J.W. Weber, M. Juhl, et al., Calculation of quantitative shunt values using photoluminescence imaging, Progress in Photovoltaics: Research and Applications 21 (5) (August 2013) 933e941. Available from: http:// onlinelibrary.wiley.com/doi/10.1002/pip.2180/ abstract. S. Herlufsen, K. Bothe, R. Brendel, J. Schmidt, Dynamic photoluminescence lifetime imaging for injection-dependent lifetime measurements, Energy Procedia 55 (Supplement C) (January 2014) 77e84. Available from: http://www.sciencedirect.com/science/ article/pii/S1876610214013058. J.A. Giesecke, D. Walter, F. Kopp, P. Rosenits, M.C. Schubert, W. Warta, Simultaneous determination of carrier lifetime and net dopant concentration of silicon wafers from photoluminescence, in: 2010 35th IEEE Photovoltaic Specialists Conference, 2010, pp. 000847e000851. M. Juhl, T. Trupke, Y. Augarten, Emitter sheet resistance from photoluminescence images, in: 2013 IEEE 39th Photovoltaic Specialists Conference (PVSC), 2013, pp. 0198e0202. J.C. Schlothauer, K. Grabmayer, I. Hintersteiner, G.M. Wallner, B. Rder, Non-destructive 2Dluminescence detection of EVA in aged PV modules: correlation to calorimetric properties, additive distribution and a clue to aging parameters, Solar Energy Materials and Solar Cells 159 (Supplement C) (January 2017) 307e317. Available from: http://www.sciencedirect.com/ science/article/pii/S092702481630349X. Y. Voronko, T. Koch, K. Berger, G. Oreski, M. Knausz, M. Weiss, et al., Long term performance of PV modules: system optimization through the application of innovative nondestructive characterization methods, in: 27th European Photovoltaic Solar Energy Conference and Exhibition, October 2012, pp. 3530e3535. Available from: http://www.eupvsec-proceedings. com/proceedings?paper¼16149. K. Ramspeck, K. Bothe, D. Hinken, B. Fischer, J. Schmidt, R. Brendel, Recombination current

4: D EGRADATION P ROCESSES

[43]

[44]

[45]

[46]

[47]

[48]

IN

P HOTOVOLTAIC C ELLS

and series resistance imaging of solar cells by combined luminescence and lock-in thermography, Applied Physics Letters 90 (15) (April 2007) 153502. Available from: http://aip. scitation.org/doi/abs/10.1063/1.2721138. M. Kasemann, D. Grote, B. Walter, W. Kwapil, T. Trupke, Y. Augarten, et al., Luminescence imaging for the detection of shunts on silicon solar cells, Progress in Photovoltaics: Research and Applications 16 (4) (June 2008) 297e305. Available from: http://onlinelibrary.wiley.com/ doi/10.1002/pip.812/abstract. S. Pingel, O. Frank, M. Winkler, S. Daryan, T. Geipel, H. Hoehne, et al., Potential induced degradation of solar cells and panels, in: Photovoltaic Specialists Conference (PVSC), 2010 35th IEEE, IEEE, 2010, pp. 002817e002822. Available from: http://ieeexplore.ieee.org/ abstract/document/5616823/. J. Berghold, O. Frank, H. Hoehne, S. Pingel, B. Richardson, M. Winkler, Potential induced degradation of solar cells and panels, in: 25th EUPVSEC, 2010, pp. 3753e3759. Available from: http://www.solon.com/export/ sites/default/solonse.com/_downloads/global/ presse/solon-fronius-valencia-PID.pdf. S. Tatapudi, F. Ebneali, J. Kuitche, G. Tamizhmani, Potential induced degradation of pre-stressed photovoltaic modules: effect of glass surface conductivity disruption, in: Photovoltaic Specialists Conference (PVSC), 2013 IEEE 39th, IEEE, 2013, pp. 1604e1609. Available from: http://ieeexplore.ieee.org/ abstract/document/6744452/. F. Ebneali, S. Tatapudi, G. TamizhMani, Potential induced degradation of pre-stressed photovoltaic modules: influence of polarity, surface conductivity and temperature, in: Photovoltaic Specialists Conference (PVSC), 2013 IEEE 39th, IEEE, 2013, pp. 1548e1553. Available from: http://ieeexplore.ieee.org/ abstract/document/6744440/. S. Goranti, G. TamizhMani, Potential induced degradation (PID) study on accelerated stress tested PV modules, in: Photovoltaic Specialists Conference (PVSC), 2012 38th IEEE, IEEE, 2012, pp. 002438e002441. Available from: http://ieeexplore.ieee.org/ abstract/document/6318088/.

113

[49] P. Hacke, K. Terwilliger, R. Smith, S. Glick, J. Pankow, M. Kempe, et al., System voltage potential-induced degradation mechanisms in PV modules and methods for test, in: Photovoltaic Specialists Conference (PVSC), 2011 37th IEEE, IEEE, 2011, pp. 000814e000820. Available from: http://ieeexplore.ieee.org/ abstract/document/6186079/. [50] S. Hoffmann, M. Koehl, Effect of humidity and temperature on the potential-induced degradation, Progress in Photovoltaics: Research and Applications 22 (2) (2014) 173e179. Available from: http://onlinelibrary.wiley.com/doi/10. 1002/pip.2238/full. [51] J. Oh, S. Bowden, G. TamizhMani, Potentialinduced degradation (PID): incomplete recovery of shunt resistance and quantum efficiency losses, IEEE Journal of Photovoltaics 5 (6) (2015) 1540e1548. Available from: http:// ieeexplore.ieee.org/abstract/document/7182748/. [52] C. Buerhop, T. Pickel, T. Blumberg, J. Adams, S. Wrana, M. Dalsass, et al., IR-images of PVmodules with potential induced degradation (PID) correlated to monitored string power output, in: SPIE Optics þ Photonics for Sustainable Energy, International Society for Optics and Photonics, 2016, 99380Je99380J. Available from: http://proceedings.spiedigitallibrary.org/ proceeding.aspx?articleid¼2559139. [53] J. Carolus, J. Govaerts, E. Voroshazi, W. De Ceuninck, M. Daenen, Voltage Dependence of Potential-Induced Degradation and Recovery on Photovoltaic One-Cell Laminates, 2017. Available from: https://uhdspace.uhasselt.be/ dspace/retrieve/22881/20170210_CarolusJ_ EUPVSEC.pdf. [54] M. Schtze, M. Junghnel, M.B. Koentopp, S. Cwikla, S. Friedrich, J.W. Mller, et al., Laboratory study of potential induced degradation of silicon photovoltaic modules, in: Photovoltaic Specialists Conference (PVSC), 2011 37th IEEE, IEEE, 2011, pp. 000821e000826. Available from: http://ieeexplore.ieee.org/abstract/ document/6186080/. [55] P. Hacke, K. Terwilliger, S. Kurtz, In-Situ Measurement of Crystalline Silicon Modules Undergoing Potential-Induced Degradation in Damp Heat Stress Testing for Estimation of Low-Light Power Performance, National

114

[56]

[57]

[58]

[59]

[60]

[61]

[62]

[63]

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Renewable Energy Laboratory (NREL), Golden, CO, 2013. Available from: https:// www.osti.gov/scitech/biblio/1090972. M.B. Koentopp, M. Krober, C. Taubitz, Toward a PID test standard: understanding and modeling of laboratory tests and field progression, IEEE Journal of Photovoltaics 6 (1) (January 2016) 252e257. W. Luo, Y.S. Khoo, P. Hacke, V. Naumann, D. Lausch, S.P. Harvey, et al., Potential-induced degradation in photovoltaic modules: a critical review, Energy and Environmental Science 10 (1) (2017) 43e68. Available from: http://pubs.rsc. org/-/content/articlehtml/2017/ee/c6ee02271e. S. Yamaguchi, K. Ohdaira, Degradation behavior of crystalline silicon solar cells in a cell-level potential-induced degradation test, Solar Energy 155 (2017) 739e744. Available from: http://www.sciencedirect.com/science/ article/pii/S0038092X17305844. R. Swanson, M. Cudzinovic, D. DeCeuster, V. Desai, J. Ju¨rgens, N. Kaminar, et al., The surface polarization effect in high-efficiency silicon solar cells, in: 15th PVSEC. Shanghai, China, 2005. P. Hacke, R. Smith, K. Terwilliger, S. Glick, D. Jordan, S. Johnston, et al., Acceleration factor determination for potential-induced degradation in crystalline silicon PV modules, in: Reliability Physics Symposium (IRPS), 2013 IEEE International, IEEE, 2013, 4Be1. 00001. Available from: http://ieeexplore.ieee. org/xpls/abs_all.jsp?arnumber¼6532009. P. Hacke, R. Smith, K. Terwilliger, G. Perrin, B. Sekulic, S. Kurtz, Development of an IEC test for crystalline silicon modules to qualify their resistance to system voltage stress, Progress in Photovoltaics: Research and Applications 22 (7) (2014) 775e783. Available from: http://onlinelibrary.wiley.com/doi/10. 1002/pip.2434/full. K.L. Wong, What is wrong with the existing reliability prediction methods? Quality and Reliability Engineering International 6 (4) (1990) 251e257. J. Hattendorf, R. Loew, W.M. Gnehr, L. Wulff, M.C. Koekten, D. Koshnicharov, et al., Potential induced degradation in mono-crystalline silicon based modules: an acceleration model, in: Proceedings of the 27th European PV Solar

AND

OTHER M ATERIALS

[64]

[65]

[66]

[67]

[68]

[69]

[70]

IN

P HOTOVOLTAIC M ODULES

Energy Conference, Frankfurt, Germany, 2012, pp. 3405e3410. C. Taubitz, M. Krber, M. Schtze, M.B. Koentopp, Kinetic description and modeling of potential induced degradation, in: Proceedings of the 28th European PV Solar Energy Conference, Paris, France, 2013, pp. 3321e3323. Available from: http://www. q-cells.co.uk/uploads/tx_abdownloads/files/ 28thEUPVSEC_2013_Taubitz_et_al_PID_01.PDF. V. Naumann, T. Geppert, S. Groer, D. Wichmann, H.J. Krokoszinski, M. Werner, et al., Potentialinduced degradation at interdigitated back contact solar cells, Energy Procedia 55 (Suppl. C) (January 2014) 498e503. Available from: http://www.sciencedirect.com/science/article/ pii/S1876610214012399. W. Oh, J. Kim, B. Kang, S. Bae, K.D. Lee, H.S. Lee, et al., Evaluation of potentialinduced degradation in crystalline Si solar cells using Na fault injection, Microelectronics Reliability 64 (2016) 646e649. Available from: http://www.sciencedirect.com/science/ article/pii/S0026271416302037. V. Naumann, D. Lausch, S. Grosser, M. Werner, S. Swatek, C. Hagendorf, et al., Microstructural analysis of crystal defects leading to potentialinduced degradation (PID) of Si solar cells, Energy Procedia 33 (2013) 76e83. Available from: http://www.sciencedirect.com/science/ article/pii/S1876610213000520. J. Bauer, V. Naumann, S. Grosser, C. Hagendorf, M. Schtze, O. Breitenstein, On the mechanism of potential-induced degradation in crystalline silicon solar cells, Physica Status Solidi (RRL)-Rapid Research Letters 6 (8) (2012) 331e333. Available from: http://onlinelibrary.wiley.com/doi/10.1002/ pssr.201206276/full. V. Naumann, D. Lausch, A. Hhnel, J. Bauer, O. Breitenstein, A. Graff, et al., Explanation of potential-induced degradation of the shunting type by Na decoration of stacking faults in Si solar cells, Solar Energy Materials and Solar Cells 120 (2014) 383e389. Available from: http://www.sciencedirect. com/science/article/pii/ S0927024813003000. D. Lausch, V. Naumann, A. Graff, A. Hhnel, O. Breitenstein, C. Hagendorf, et al., Sodium

4: D EGRADATION P ROCESSES

[71]

[72]

[73]

[74]

[75]

[76]

[77]

IN

P HOTOVOLTAIC C ELLS

outdiffusion from stacking faults as root cause for the recovery process of potentialinduced degradation (PID), Energy Procedia 55 (2014) 486e493. Available from: http:// www.sciencedirect.com/science/article/pii/ S1876610214012375. J. Kapur, K.M. Stika, C.S. Westphal, J.L. Norwood, B. Hamzavytehrany, Prevention of potential-induced degradation with thin ionomer film, IEEE Journal of Photovoltaics 5 (1) (2015) 219e223. Available from: http:// ieeexplore.ieee.org/abstract/document/6963261/. R. Tanaka, H. Zenkoh, Encapsulant effect on PID durability of various crystalline PV cells, in: Photovoltaic Module Rel. Workshop, 2015. Available from: https://www.nrel.gov/pv/assets/ pdfs/2015_pvmrw_41_tanaka.pdf. M. Schwark, K. Berger, R. Ebner, G. Ujvari, C. Hirschl, L. Neumaier, et al., Investigation of potential induced degradation (PID) of solar modules from different manufacturers, in: Industrial Electronics Society, IECON 2013-39th Annual Conference of the IEEE, IEEE, 2013, pp. 8090e8097. Available from: http://ieeexplore.ieee.org/abstract/document/ 6700486/. J. Berghold, S. Koch, B. Frohmann, P. Hacke, P. Grunow, Properties of encapsulation materials and their relevance for recent field failures, in: Photovoltaic Specialist Conference (PVSC), 2014 IEEE 40th, IEEE, 2014, pp. 1987e1992. Available from: http://ieeexplore.ieee.org/ abstract/document/6925315/. M.C. Lpez-Escalante, L.J. Caballero, F. Martn, M. Gabs, A. Cuevas, J.R. Ramos-Barrado, Polyolefin as PID-resistant encapsulant material in PV modules, Solar Energy Materials and Solar Cells 144 (2016) 691e699. Available from: http://www.sciencedirect.com/science/ article/pii/S0927024815005206. C. Deline, B. Marion, J. Granata, S. Gonzalez, Performance and Economic Analysis of Distributed Power Electronics in Photovoltaic Systems, National Renewable Energy Laboratory (NREL), Golden, CO, 2011. Available from: http://www.nrel.gov/docs/fy11osti/50003.pdf. C. Deline, J. Meydbray, M. Donovan, J. Forrest, Partial shade evaluation of distributed power electronics for photovoltaic systems, in: Photovoltaic Specialists Conference (PVSC), 2012

115

[78]

[79]

[80]

[81]

[82]

[83]

[84]

[85]

38th IEEE, IEEE, 2012, pp. 001627e001632. Available from: http://ieeexplore.ieee.org/ abstract/document/6317908/. M.A. Hossain, Y. Xu, T.J. Peshek, L. Ji, A.R. Abramson, R.H. French, Microinverter thermal performance in the real-world: measurements and modeling, PLoS One 10 (7) (2015) e0131279. Available from: http:// journals.plos.org/plosone/article?id¼10.1371/ journal.pone.0131279. J. Lindroos, H. Savin, Review of lightinduced degradation in crystalline silicon solar cells, Solar Energy Materials and Solar Cells 147 (2016) 115e126. Available from: http://www.sciencedirect.com/science/article/ pii/S0927024815006406. H. Fischer, W. Pschunder, in: Proceedings of the Tenth IEEE Photovoltaic Specialists Conference, IEEE, New York, 1973. J. Schmidt, A.G. Aberle, R. Hezel, Investigation of carrier lifetime instabilities in Cz-grown silicon, in: Photovoltaic Specialists Conference, 1997, Conference Record of the Twenty-Sixth IEEE, IEEE, 1997, pp. 13e18. Available from: http://ieeexplore.ieee.org/abstract/ document/653914/. S. Rein, S.W. Glunz, Electronic properties of the metastable defect in boron-doped Czochralski silicon: unambiguous determination by advanced lifetime spectroscopy, Applied Physics Letters 82 (7) (February 2003) 1054e1056. Available from: http://aip. scitation.org/doi/10.1063/1.1544431. H. Hashigami, Y. Itakura, T. Saitoh, Effect of illumination conditions on Czochralski-grown silicon solar cell degradation, Journal of Applied Physics 93 (7) (March 2003) 4240e4245. Available from: http://aip. scitation.org/doi/10.1063/1.1559430. J. Schmidt, R. Hezel, light-induced degradation in CZ silicon solar cells: fundamental understanding and strategies for its avoidance, in: 12th Workshop on Crystalline Silicon Solar Cell Materials and Processes. Breckenridge, 2002, pp. 321e325. Available from: http:// www.ilb-helios-reh.com/bibliothek/workshop. pdf. J.H. Reiss, R.R. King, K.W. Mitchell, Characterization of diffusion length degradation in Czochralski silicon solar cells, Applied Physics

116

[86]

[87]

[88]

[89]

[90]

[91]

[92]

[93]

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Letters 68 (23) (1996) 3302e3304. Available from: http://aip.scitation.org/doi/abs/10.1063/ 1.116581. J. Schmidt, A. Cuevas, Electronic properties of light-induced recombination centers in borondoped Czochralski silicon, Journal of Applied Physics 86 (6) (1999) 3175e3180. Available from: http://aip.scitation.org/doi/abs/10.1063/ 1.371186. T. Saitoh, H. Hashigami, S. Rein, S. Glunz, Overview of light degradation research on crystalline silicon solar cells, Progress in Photovoltaics: Research and Applications 8 (5) (2000) 537e547. Available from: http:// onlinelibrary.wiley.com. S.W. Glunz, S. Rein, J. Knobloch, W. Wettling, T. Abe, Comparison of boron- and galliumdoped p-type Czochralski silicon for photovoltaic application, Progress in Photovoltaics: Research and Applications 7 (6) (1999) 463e469. Available from: https:// www.researchgate.net/profile/Stefan_Glunz/ publication/230601727_Comparison_of_Boron_ and_Gallium_doped_p-type_Czochralski_silicon_for_photovoltaic_application/links/0912f 5107c725ee810000000.pdf. T.U. Naerland, Characterization of Light Induced Degradation in Crystalline Silicon, 2013. Available from: https://brage.bibsys.no/ xmlui/handle/11249/249425. K. Bothe, J. Schmidt, Electronically activated boron-oxygen-related recombination centers in crystalline silicon, Journal of Applied Physics 99 (1) (January 2006) 013701. Available from: http://aip.scitation.org/doi/10.1063/1.2140584. J. Adey, R. Jones, D.W. Palmer, P.R. Briddon, S. berg, Degradation of boron-doped Czochralski-grown silicon solar cells, Physical Review Letters 93 (5) (2004) 055504. Available from: https://journals.aps.org/prl/abstract/10. 1103/PhysRevLett.93.055504. K. Bothe, R. Sinton, J. Schmidt, Fundamental boronoxygen-related carrier lifetime limit in mono-and multicrystalline silicon, Progress in Photovoltaics: Research and Applications 13 (4) (2005) 287e296. Available from: http:// onlinelibrary.wiley.com/doi/10.1002/pip.586/full. J. Schmidt, A. Cuevas, S. Rein, S.W. Glunz, Impact of light-induced recombination centres on the current-voltage characteristic of

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Czochralski silicon solar cells, Progress in Photovoltaics: Research and Applications 9 (4) (2001) 249e255. Available from: http:// onlinelibrary.wiley.com/doi/10.1002/pip.373/full. [94] J. Schmidt, K. Bothe, Structure and transformation of the metastable boron-and oxygenrelated defect center in crystalline silicon, Physical Review B 69 (2) (2004) 024107. Available from: https://journals.aps.org/prb/ abstract/10.1103/PhysRevB.69.024107. [95] L.I. Murin, E.A. Tolkacheva, V.P. Markevich, A.R. Peaker, B. Hamilton, E. Monakhov, et al., The oxygen dimer in Si: its relationship to the light-induced degradation of Si solar cells? Applied Physics Letters 98 (18) (2011) 182101. Available from: http://aip.scitation.org/doi/abs/ 10.1063/1.3584138. [96] V.V. Voronkov, R. Falster, K. Bothe, B. Lim, J. Schmidt, Lifetime-degrading boron-oxygen centres in p-type and n-type compensated silicon, Journal of Applied Physics 110 (6) (September 2011) 063515. Available from: http://aip.scitation.org/doi/10.1063/1.3609069. [97] V.V. Voronkov, R. Falster, K. Bothe, B. Lim, Light-induced lifetime degradation in borondoped Czochralski silicon: are oxygen dimers involved? Energy Procedia 38 (Suppl. C) (January 2013) 636e641. Available from: http://www.sciencedirect.com/science/article/ pii/S1876610213014136. [98] V.V. Voronkov, R. Falster, Latent complexes of interstitial boron and oxygen dimers as a reason for degradation of silicon-based solar cells, Journal of Applied Physics 107 (5) (March 2010) 053509. Available from: http://aip. scitation.org/doi/10.1063/1.3309869. [99] B. Hallam, M. Abbott, J. Bilbao, P. Hamer, N. Gorman, M. Kim, et al., Modelling kinetics of the boron-oxygen defect system, Energy Procedia 92 (2016) 42e51. Available from: http://www.sciencedirect.com/science/article/ pii/S1876610216304350. [100] P. Hamer, N. Nampalli, Z. Hameiri, M. Kim, D. Chen, N. Gorman, et al., Boron-oxygen defect formation rates and activity at elevated temperatures, Energy Procedia 92 (2016) 791e800. Available from: http://www.sciencedirect.com/ science/article/pii/S1876610216304970. [101] J.E. Cotter, J.H. Guo, P.J. Cousins, M.D. Abbott, F.W. Chen, K.C. Fisher, P-type

4: D EGRADATION P ROCESSES

[102]

[103]

[104]

[105]

[106]

[107]

[108]

IN

P HOTOVOLTAIC C ELLS

versus n-type silicon wafers: prospects for high-efficiency commercial silicon solar cells, IEEE Transactions on Electron Devices 53 (8) (August 2006) 1893e1901. Available from: http://ieeexplore.ieee.org/document/1661892/. M. Xie, C. Ren, L. Fu, X. Qiu, X. Yu, D. Yang, An industrial solution to light-induced degradation of crystalline silicon solar cells, Frontiers in Energy 11 (1) (2017) 67e71. Available from: http://link.springer.com/article/10.1007/ s11708-016-0430-x. P. Grunow, P. Clemens, V. Hoffmann, B. Litzenburger, L. Podlowski, Influence of Micro Cracks in Multi-Crystalline Silicon Solar Cells on the Reliability of PV Modules. Barcelona, Spain, 2005. Available from: http:// www.solon.com/export/sites/default/solonse. com/_downloads/global/article-pid/Grunow_ et_al_Influence_of_Micro_Cracks.pdf. S. Pingel, Y. Zemen, O. Frank, T. Geipel, J. Berghold, Mechanical stability of solar cells within solar panels, in: Proc 24th European Photovoltaic Energy Conf, January 2009. J.H. Wohlgemuth, D.W. Cunningham, N.V. Placer, G.J. Kelly, A.M. Nguyen, The effect of cell thickness on module reliability, in: 2008 33rd IEEE Photovoltaic Specialists Conference, 2008, pp. 1e4. C.W. Hardin, J. Qu, A.J. Shih, Fixed abrasive diamond wire saw slicing of single-crystal silicon carbide wafers, Materials and Manufacturing Processes 19 (2) (December 2004) 355e367. https://doi.org/10.1081/AMP120029960. Available from:. A.S. Azar, B. Holme, y Nielsen, Effect of sawing induced micro-crack orientations on fracture properties of silicon wafers, Engineering Fracture Mechanics 154 (Suppl. C) (March 2016) 262e271. Available from: http://www.sciencedirect.com/science/article/ pii/S0013794416000175. B.k Sultan, E. Srvik, Thermal degradation of EVA and EBAA comparison. I. Volatile decomposition products, Journal of Applied Polymer Science 43 (9) (November 1991) 1737e1745. Available from: http://onlinelibrary. wiley.com/doi/10.1002/app.1991.070430917/ abstract.

117

[109] M. Kontges, F. Wegert, R. Illing, M. Wegner, M. Denz, A. de la Dedicacin Rodrguez, et al., Impact of transportation on silicon wafer-based PV modules, in: 28th European Photovoltaic Solar Energy Conference and Exhibition, November 2013, pp. 2960e2967. Available from: http://www.eupvsec-proceedings.com/ proceedings?paper¼24495. [110] F. Reil, K. Strohkfendl, W. Herrmann, W. Vaassen, J. Althaus, The effect of transportation impacts and dynamic load tests on the mechanical and electrical behaviour of crystalline PV modules, in: 25th European Photovoltaic Solar Energy Conference and Exhibition/ 5th World Conference on Photovoltaic Energy Conversion, 6e10 September 2010, Valencia, Spain, October 2010, pp. 3989e3992. Available from: http://www.eupvsec-proceedings.com/ proceedings?paper¼6678. [111] M. Assmus, S. Jack, K.A. Weiss, M. Koehl, Measurement and simulation of vibrations of PV-modules induced by dynamic mechanical loads, Progress in Photovoltaics 19 (Nr.6) (2011) 688e694. [112] S. Kajari-Schrder, I. Kunze, U. Eitner, M. Kntges, Spatial and orientational distribution of cracks in crystalline photovoltaic modules generated by mechanical load tests, Solar Energy Materials and Solar Cells 95 (11) (November 2011) 3054e3059. Available from: http://www.sciencedirect.com/science/article/ pii/S0927024811003710. [113] H. Xiong, C. Gan, X. Yang, Z. Hu, H. Niu, J. Li, et al., Corrosion behavior of crystalline silicon solar cells, Microelectronics Reliability 70 (Suppl. C) (March 2017) 49e58. Available from: http://www.sciencedirect.com/science/ article/pii/S0026271417300069. [114] C. Peike, S. Hoffmann, P. Hlsmann, K.A. Wei, M. Koehl, P. Bentz, Permeation Impact on Metallization Degradation, 2012, p. 84720U1e84720U-9. Available from: http:// proceedings.spiedigitallibrary.org/proceeding. aspx?articleid¼1379672. [115] P. Snchez-Friera, M. Piliougine, J. Pelez, J. Carretero, M. Sidrach de Cardona, Analysis of degradation mechanisms of crystalline silicon PV modules after 12 years of

118

[116]

[117]

[118]

[119]

[120]

D URABILITY AND R ELIABILITY

OF

P OLYMERS

operation in Southern Europe, Progress in Photovoltaics: Research and Applications 19 (6) (September 2011) 658e666. Available from: http://onlinelibrary.wiley.com/doi/10. 1002/pip.1083/abstract. A. Kraft, L. Labusch, T. Ensslen, I. Drr, J. Bartsch, M. Glatthaar, et al., Investigation of acetic acid corrosion impact on printed solar cell contacts, IEEE Journal of Photovoltaics 5 (3) (May 2015) 736e743. G. Oreski, A. Mihaljevic, Y. Voronko, G.C. Eder, Acetic acid permeation through photovoltaic backsheets: influence of the composition on the permeation rate, Polymer Testing 60 (Suppl. C) (July 2017) 374e380. Available from: http://www.sciencedirect.com/ science/article/pii/S014294181730209X. M.C.C.D. Oliveira, A.S.A. Cardoso Diniz, M.M. Viana, V.D.F.C. Lins, The causes and effects of degradation of encapsulant ethylene vinyl acetate copolymer (EVA) in crystalline silicon photovoltaic modules: a review, Renewable and Sustainable Energy Reviews (July 2017) 2299e2317. Available from: http:// www.sciencedirect.com/science/article/pii/ S1364032117309851. T.H. Kim, N.C. Park, D.H. Kim, The effect of moisture on the degradation mechanism of multi-crystalline silicon photovoltaic module, Microelectronics Reliability 53 (9) (September 2013) 1823e1827. Available from: http:// www.sciencedirect.com/science/article/pii/ S0026271413002229. E. Wang, H.E. Yang, J. Yen, S. Chi, C. Wang, Failure modes evaluation of PV module via materials degradation approach, Energy Procedia 33 (Suppl. C) (January 2013) 256e264.

AND

OTHER M ATERIALS

[121]

[122]

[123]

[124]

[125]

IN

P HOTOVOLTAIC M ODULES

Available from: http://www.sciencedirect.com/ science/article/pii/S1876610213000763. C. Peike, S. Hoffmann, P. Hlsmann, B. Thaidigsmann, K.A. Wei, M. Koehl, et al., Origin of damp-heat induced cell degradation, Solar Energy Materials and Solar Cells 116 (Suppl. C) (September 2013) 49e54. Available from: http://www.sciencedirect.com/science/ article/pii/S0927024813001359. Y. Zhang, J. Xu, J. Mao, J. Tao, H. Shen, Y. Chen, et al., Long-term reliability of silicon wafer-based traditional backsheet modules and double glass modules, RSC Advances 5 (81) (July 2015) 65768e65774. Available from: http://pubs.rsc.org/en/content/articlelanding/ 2015/ra/c5ra11224a. IEC, IEC 61215 Terrestrial Photovoltaic (PV) Modules - Design Qualification and Type Approval, International Electrotechnical Commission, 2016. Available from: https:// webstore.iec.ch/publication/24312. A.W. Czanderna, F.J. Pern, Encapsulation of PV modules using ethylene vinyl acetate copolymer as a pottant: a critical review, Solar Energy Materials and Solar Cells 43 (2) (September 1996) 101e181. Available from: http://www.sciencedirect.com/science/article/pii/ 0927024895001506. M.D. Kempe, G.J. Jorgensen, K.M. Terwilliger, T.J. McMahon, C.E. Kennedy, T.T. Borek, Acetic acid production and glass transition concerns with ethylene-vinyl acetate used in photovoltaic devices, Solar Energy Materials and Solar Cells 91 (4) (February 2007) 315e329. Available from: http://www.sciencedirect.com/ science/article/pii/S0927024806004107.

5 Degradation and Failure Mechanisms of PV Module Interconnects Yang Hu 1 and Roger H. French 2 GE Renewable Energy, San Ramon, CA, United States 2 Case Western Reserve University, Case School of Engineering, SDLE Research Center, Materials Science and Engineering, Cleveland, OH, United States 1

5.1 Introduction to PV Modules Interconnection, Degradation, and Failures The crystalline silicon (c-Si) wafer-based solar cells have been successfully transitioned from the research laboratory to mass production to become the dominant technology in commercialized solar photovoltaic (PV) modules [1]. In this chapter, we will focus on degradation and failure mechanisms of the c-Si cell interconnections in commercialized c-Si PV modules. The screen printed silver (Ag) front contact aluminum (Al) back surface field (Al-BSF) cell has been the dominant commercial solar cell architecture. But the Al-BSF architecture is now being superseded by the passivated emitter rear contact cell (PERC), which is similar from an interconnection perspective. The assembly and manufacturing process of front Ag contact silicon solar cells involves screen printing the front metal electrode and the back aluminum contacts on the cell using fritted glass paste materials, followed by soldering highly conductive lead/tin solder-coated ribbon along the front-side silver busbar [2]. The ribbon extends from 1 cell to the next and is soldered to the back of a neighboring cell to enable the current flow from the front side of one cell to the back side of the next cell in a series connection [3]. The manufacturing process of PV module interconnections involves the use of infrared reflow soldering, induction soldering, or even hand soldering. The soldered joint provides good electrical contact between the solar cells and electrodes, but it can induce high mechanical stress and thermal shock which can induce microcracking in the cells. The stringing ribbon connects solar cells with each other to form strings, while the bussing ribbon connects

strings of solar cells to form a module. The bus ribbon is soldered inside junction box with bypass diodes to protect the module from hot spot and shading effects. The leads coming out of the junction box are the output interface of the PV module. The failures of cell interconnection in c-Si PV modules have been reported as a key reliability challenge [3e6]. The interconnect ribbon is a wide and flat-shaped copper (Cu) metal wire soldered by tin-lead-silver (SnPbAg) on the front side of one PV cell and the back side of neighboring PV cell, as shown in Fig. 5.1. Metallic corrosion, induced by hygrothermal stress on screen-printed silver metal busbars or grid lines, is well known to lead to power degradation of c-Si PV modules [7e9]. The degradation of solder joints during the module’s field operation due to temperature cycling has been

Figure 5.1 Solder interconnection between ribbon wire and silicon solar cell. From J.S. Jeong, N. Park, C. Han, Field failure mechanism study of solder interconnection for crystalline silicon photovoltaic module, Microelectronics Reliability 52(9e10) (2012) 2326e2330.

Durability and Reliability of Polymers and Other Materials in Photovoltaic Modules. https://doi.org/10.1016/B978-0-12-811545-9.00005-7 Copyright © 2019 Yang Hu & Roger H. French. Published by Elsevier Inc. All rights reserved.

119

120

D URABILITY AND R ELIABILITY

OF

P OLYMERS

reported as one major contributor to PV module’s interconnection failures. Extreme solder cracks can result in the PV module’s open circuit failure [6,10,11]. In a PV module, all the cells and their interconnections are encapsulated between the glass front sheet and the polymeric backsheet. The difference in thermal expansion coefficients of the ribbon and silicon results in the variation of cell-to-cell spacing at different temperatures. Thus, stress can accumulate in the interconnect ribbon, and especially the part between two cells where the ribbon adopts the sigmoid shape of the front surface to back surface “tab.” Furthermore, stress from the interconnection can be concentrated at the edge of solar cell where the ribbon is bent. As a result, interconnect ribbon failure originating from metal fatigue is a common failure mode in flat-plate PV modules [12e14]. The degradation of the solder joint at the electrical connection of the string bypass diode in the junction box may also be induced by repeated thermal cycling over lifetime. Severe heat damage, even up to PV electrical fires, has occurred at junction box and may be exacerbated by DC arcing at the interconnection crack caused by solder joint fatigue. We will describe here the three typical interconnection degradation and failure mechanisms of silicon as solder: front-side silver grid corrosion, solder joint degradation, and interconnect ribbon fatigue. In addition, there has been increased recent interest in replacing the copper ribbon interconnect with electrically conductive adhesives (ECAs), used with PV cells that overlap each other (referred to as “shingled” cells), and the reliability issues of these will also be discussed.

5.2 Front-Side Silver Grid Corrosion The standard front-side metallization of crystalline silicon solar cell is done by screen printing silver containing paste on solar cell’s front surface and heat treated to form an ohmic contact [2]. In order to minimize the shading of metallization on the front surface, the silver gridline width is less than 100 mm after firing [15,16]. Under accelerated reliability tests and long-term outdoor exposure, front-side metallization is a known weakness of PV modules [17e20]. The damp heat (DH) test, at 85
C and 85% relative humidity [21], is a common and crucial test for PV

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 5.2 Electroluminescence images of photovoltaic modules exposed in 85
C/85% relative humidity (RH) test for 3000 h.

modules qualification testing. Because of the high humidity during the DH test, polymeric and metallic compounds of PV modules may degrade, inducing discoloration, delamination, or corrosion [22e24]. A frequently observed degradation phenomenon after DH test is the dark outer area of each solar cell in the electroluminescence (EL) image [20]. This DHinduced degradation (DHID) is typically accompanied by serious decreases in module performance. A similar dark region is detected in PV modules that went through long outdoor exposures. Fig. 5.2 shows the EL image of a PV module subjected to exposure in 85
C and 85% relative humidity test conditions for 3000 h. The dark regions along the edge of each solar cell were observed. Although the EL dark area’s distribution depends on the architecture of PV modules, the root cause of such degradation is moisture ingress into PV module and acetic acid (HAc) generated from the hydrolysis decomposition of ethylene vinyl acetate (EVA) [20]. The corrosion of front-side metallization corrosion is primarily caused by a high concentration of HAc. The acetic acid attacks the silver gridline and solar cell interface, which is the critical current path for PV electronic current.

5.2.1 Mechanisms The front-side Ag grid is screen printed with an Ag paste containing fritted (ground) glass particles. Ag paste contains Ag particles of an average size of 1 mm (80%e90%), glass frit (0%e5%), an organic solvent (3%e15%), a cellulose resin (3%e15%), and inorganic additives and surfactants (1%e2%) [25]. The additives are supposed to lower the co-fire temperature, help minimize the shrinkage mismatch with the

5: D EGRADATION

AND

FAILURE M ECHANISMS

OF

PV M ODULE I NTERCONNECTS

dielectric, and increase mechanical strength. The glass frit dissolves the antireflective coating on the front surface of the cell and makes ohmic contact between the Ag paste and the pen junction. Peike et al. showed in 2013 that the failure mechanism of DHID is due to a loss of the electrical contact between emitter and front-side metallization [20]. They suggested that DHID originates presumably from an attack of HAc on the glass frit inside the front-side contact. As the current standard module encapsulation material of c-Si solar cells, EVA is known to degrade under the presence of humidity [26]. HAc is a by-product of the EVA hydrolysis. The currentevoltage curve measurement of the aged PV module showed a decreased series fill factor (FF) caused by an increased series resistance. Moreover, the spatial distribution of the series resistance (Rs) calculated from luminescent images showed a significantly increased series resistance in the EL dark region. Therefore, grid corrosion or a reduced conductivity between the emitter and the grid is the most likely cause of the DHID [20]. Kraft et al. demonstrated the dissolution of the glass layer underneath the silver gridline metallization by immersing solar cells into aqueous HAc solution. In standard lead glass, lead oxide is present, which is observed to corrode in the presence of HAc. According to the Pourbaix diagrams, the following reaction is probable: PbO þ 2CH3 COOH/ðCH3 COOÞ2Pb þ H2 O. (5.1)

121

Lead oxide reacts in the presence of acetic acid to lead acetate, which is highly soluble in water. Water transports into the modules through the backsheet, diffuses at the cell front side from the edges to the center of the cell and moves the HAc present within the EVA encapsulation to the contacts [17,20]. Moisture transports more acetic acid into the vicinity of the glassesilver interface and removes the lead acetate compound. The corrosion process can be accelerated by negative voltage applied to the front contact even at low HAc concentration. This behavior can be explained by the reduction of the dissolved lead in the presence of electrons at a suitable potential, which will remove the lead from the system and release the HAc. This electrochemically driven dissolution cycle is shown in Fig. 5.3 as a schematic drawing at the contact interface. The reaction is particularly strong where the transport distance between the place of lead dissolution (glass layer) and the place of lead reduction (bulk silver) is very short. This leads to a quick spreading gap formation between the glass and silver inside the contact. This is the origin of the adhesion loss due to the presence of acetic acid. The dissolution mechanism explains the observation that high-series resistance cell area is highly correlated to the dark area in EL images. The kinetic analysis for PV and electrical characteristics was conducted experimentally by detecting the electrical signal concerned in contact gap formation by HAc vapor. This signal seems to be a crucial “aging signature” in PV modules.

Figure 5.3 A schematic drawing of the dissolution mechanism at the glass silver boundary layer in the contact. This mechanism causes poor contact adhesion and works efficiently if a voltage is applied to the contacts, due to the redeposition of the dissolved species. From A. Kraft, L. Labusch, T. Ensslen, I, Du¨rr, J. Bartsch, M. Glatthaar, et al., Investigation of acetic acid corrosion impact on printed solar cell contacts. IEEE Journal of Photovoltaics 5(3) (2015) 736e743.

122

D URABILITY AND R ELIABILITY

OF

P OLYMERS

5.2.2 Characterization The gap evolved by the dissolution of the glass layer underneath the silver grid lines of the solar cell may be monitored by the alterations of capacitance and/or impedance under alternating current loading conditions (Fig. 5.4). Experimentally, Tanahashi et al. mounted PV cells soldered with cell interconnect ribbons in a chromatography chamber filled with HAc vapor under high humidity and high temperature (Fig. 5.5). This exposure system was constructed by reference to Kempe’s previous work [26]. PV minimodules (c.4 W, 180  180 mm) are assembled with the same type of PV cell plus extra EVA, backsheet, and glass. The DH test of PV minimodules was carried out at 85
C/85% relative humidity over 3000 h. The characteristics of AC impedance were evaluated by an LCR meter with frequency scanning function [28]. The degradation profile of minimodules is divided to two phases by the electrical characteristics. First, the generating power (Pmax) was rapidly declined with decreasing in FF after a short-time lag (phase I), and thereafter the gradual reduction of Pmax with that of Isc (phase II) was observed (Fig. 5.6). It is worth noting that at phase I, the development of a new capacitance C3 with higher capacitance than C2

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

(capacitance derived from pen junction) was identified by AC impedance spectroscopy (Fig. 5.6). Simultaneously, the increase of series resistance, RS, in DC and R1 in AC measurement was identified by analysis for dark IeV curve and AC impedance spectrum. The resistance R3, which is developed with the new capacitance C3, increased through phase I. The degradation evolution is also observed through the evolution of EL images (Fig. 5.7). Bright “cloud”-like area and bright points semi-uniformly distributed in the lower EL brightness background were observed in the initial EL image of the sample cell. The bright area clears away during phase I and is replaced by sparsely distributed bright points only on finger electrodes under dark EL background. This result suggests that the gap underneath the finger electrodes developed and grew during phase I (shown in Fig. 5.8). However, the direct contact of silver and emitter surface still appears to remain. These contact points are shown as an Ag pillar in Fig. 5.8. In phase II, the series resistance (Rs and R1), the novel impedance-derived resistance R3, and FF stay nearly constant (Fig. 5.6). However, Pmax and Isc decreased during the whole duration of phase II. It should be noted that Isc decreases as the EL image “cloud”-like brightness disappears, and EL brightness was limited to the area near the busbar on the PV

Figure 5.4 Formation of the gap underneath the finger electrode on a p-type c-Si photovoltaic cell by corrosion, and the respective AC equivalent circuits modeled under dark conditions. Intact contact between the front electrode (orange) (dark gray in print versions) and the emitter of the Si wafer (yellow) (light gray in print versions) is illustrated in (A), and the corroded contact with formed gap is demonstrated in (B). From T. Tanahashi, N. Sakamoto, H. Shibata, A. Masuda, Localization and characterization of a degraded site in crystalline silicon photovoltaic cells exposed to acetic acid vapor. IEEE Journal of Photovoltaics 8(4) 2018 997e1004. https://doi.org/10. 1109/JPHOTOV.2018.2839259.

5: D EGRADATION

AND

FAILURE M ECHANISMS

OF

PV M ODULE I NTERCONNECTS

123

Figure 5.5 Experimental setup for the exposure of bare photovoltaic (PV) cells to acetic acid (HAc) vapor. From N. Tanahashi, H. Sakamoto, A. Shibata, Masuda, Localization and characterization of a degraded site in crystalline silicon photovoltaic cells exposed to acetic acid vapor. IEEE Journal of Photovoltaics 8(4) (2018) 997e1004. https://doi.org/10. 1109/JPHOTOV.2018.2839259.

cell. From this observation, it is assumed that some alteration with addition of forward directional diode and/or increase of resistivity occurs at the Ag pillars (Fig. 5.8). The current pass between emitter and silver finger consists of two routes: direct contact via Ag pillars and electron tunneling via nano-Ag colloids dispersed in glass layer [30,31]. Therefore, it is assumed that the FF reduction during phase I is due to dissolution of glass layer that contains dispersed nanosilver particles. In phase II, the PV cells begin to lose the diode blocking characteristic in reverse bias after 24 h of exposure to HAc vapor and 85
C/85% relative humidity. This result together with the Isc decrease suggests the electrical property of the Ag pillars have changed. Similar progression in phases of IeV curve and AC impedance parameters were observed in the PV cell encapsulated in a conventional module architecture (Fig. 5.9). The ingress of moisture onto PV cells within a module progresses from the periphery to the central area in a constant humidity damp heat test [32]. The

degradation of a PV cell caused by moisture starts from the periphery area first, and then subsequently occurs in the central area of PV cell under DH test [20,33]. Because nonuniform degradation phases can occur within one PV cell, the electrical signal from each region in different degradation phases may combine together and lose their characteristic features when the signal is obtained from a whole PV cell. An example of this feature loss is described by the development of impedance C3 and R3 where the increase of impedance in areas at degradation phase I is combined with the existing high impedance at areas at degradation phase II. However, the signal of phase transition still can be detected. Therefore, the degradation process identified on a bare PV cell exposed to HAc vapor, which seems to be spatialisotropically degraded, is able to be captured as an “aging signature” even in a PV cell laminated in a PV module, which seems to be spatial-anisotropically degraded under the DH stress. The acceleration factor between power loss observed in bare PV cells exposed to HAc vapor at 85
C/85% relative humidity (Fig. 5.6) and in PV

124

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Figure 5.6 A plot of the phase transition during degradation in photovoltaic cells exposed to acetic acid vapor. From T. Tanahashi, N. Sakamoto, H. Shibata, A. Masuda, Electrical detection of gap formation underneath finger electrodes on c-Si PV cells exposed to acetic acid vapor under hygrothermal conditions. In: 2016 IEEE 43rd Photovoltaic Specialists Conference (PVSC), 2016, pp. 1075e1079, https://doi.org/10.1109/PVSC.2016.7749778.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

module exposed to DH test conditions (Fig. 5.9) is roughly estimated as 100. The acceleration factor between PV modules’ power loss in DH conditions and in tropical climate conditions has been estimated to be 23 [34,35]. Using these estimations, 2300 times acceleration is induced between exposed bare PV cell to HAc vapor and outdoor exposure for PV modules in a tropical climate. However, there are some problems of simply applying the product of two acceleration factors. First, HAc is a by-product of EVA hydrolysis. Its concentration inside PV modules is strongly affected by the vapor transport rate of the backsheet and edge seal if present. The time required for the moisture to reach the center of the PV cell’s front side in a PV module is about 2000e3000 h even under DH stress condition [36,37]. This time is not considered when a PV cell is directly exposed to HAc vapor. Second, HAc concentration is not uniform in a PV module. Since the amount of HAc in each region within PV module depends on the concentration of moisture, the anisotropic degradation observed in PV module is not reflected in the exposure of PV cell to HAc vapor [38,39]. Finally, the combined effects of temperature and humidity on the hygrothermal degradation of PV modules have yet to be completely resolved. It has not been elucidated whether the time to failure of PV modules under hygrothermal conditions are determined by the exponential corrosion model or the power law model [35,40,41]. However, the “aging signature” detected in PV modules over a

Figure 5.7 Evolution of electroluminescence image during degradation in photovoltaic cells exposed to acetic acid vapor. From T. Tanahashi, N. Sakamoto, H. Shibata, A. Masuda, Electrical detection of gap formation underneath finger electrodes on c-Si PV cells exposed to acetic acid vapor under hygrothermal conditions. In: 2016 IEEE 43rd Photovoltaic Specialists Conference (PVSC), 2016, pp. 1075e1079, https://doi.org/10.1109/PVSC.2016.7749778.

5: D EGRADATION

AND

FAILURE M ECHANISMS

OF

PV M ODULE I NTERCONNECTS

125

Figure 5.8 A diagram of the putative degradation process on c-Si photovoltaic cells exposed to acetic acid vapor. From T. Tanahashi, N. Sakamoto, H. Shibata, A. Masuda, Electrical detection of gap formation underneath finger electrodes on c-Si PV cells exposed to acetic acid vapor under hygrothermal conditions. In: 2016 IEEE 43rd Photovoltaic Specialists Conference (PVSC), 2016, pp. 1075e1079, https://doi.org/10.1109/PVSC.2016.7749778.

long-term outdoor exposure is also easily identified in a PV cell under an accelerated test. The signature will be available as an index to estimate their degradation level/phase induced in the installed environment [42]. Especially, it is assumed that the parameters specified from AC impedance spectrum are not convoluted with other degradation mechanisms caused by light irradiation on PV cells (e.g., discoloration of encapsulant). Further research is needed to fully correlate module failure in the field to accelerated testing results.

5.3 Thermo-Putative Degradation and Mechanical Failure of Solder Joints

Figure 5.9 IeV curve and AC impedance parameter of a photovoltaic minimodule as function of the damp heat stress duration. From T. Tanahashi, N. Sakamoto, H. Shibata, A. Masuda, Electrical detection of gap formation underneath finger electrodes on c-Si PV cells exposed to acetic acid vapor under hygrothermal conditions. In: 2016 IEEE 43rd Photovoltaic Specialists Conference (PVSC), 2016, pp. 1075e1079, https://doi.org/10.1109/PVSC.2016.7749778.

Solder joint interconnects serve two important purposes: (1) form the electrical connection between the solar cell and copper ribbon and (2) form the mechanical bond that holds the copper ribbon attached to silicon cells. Thermal fatigue of solder joints that attach the module’s stringing ribbon to its solar cells is one typical mechanism of PV modules degradation and ultimate failure. The presence of cracks at the solder joint reduces the area of connection intersection, thus increasing the series resistance. Grain coarsening is also reported as evidence of solder joint thermal fatigue after long-term field exposure [24,43]. The key materials used in the PV module soldering are PbSn, and a solder joint is connecting silicon cell, Ag-based grids, and copper interconnect ribbon. The thermal fatigue problem is

126

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

critical for the solder joints reliability, due to the coefficient of thermal expansion (CTE) mismatch of the joint materials.

5.3.1 Mechanisms Ag leaching into solder and long-term solder joint fatigue are two major mechanisms that cause solder joint failures in c-Si solar cell [44]. Metals such as Ag and Cu are easily dissolved into solder. The dissolution speeds of Ag and Cu, when immersed to PbSn solder, are 10 and 0.09 mm/s at 260
C [45]. At the PnSn solder joint connected on Ag electrode, it was found that Ag dissolved into solder forms Ag3Sn compound with rigid and brittle characteristics [46]. The fatigue crack was observed at the interface of PbSn solder and the region of solder mixing with Ag3Sn after 250 thermal cycles from þ85 to 40
C [47]. The cross-sectional view of the solder joint with a crack observed by optical microscopy after 1000 thermal cycles from þ85 to 40
C is shown in Fig. 5.10 [44]. Ag electrode is used to connect to Cu ribbon interconnection by solder. There are two interfaces, Agesolder interface and Cuesolder interface. The crack shown in Fig. 5.10 occurred between the Ag and solder interface. The compound of Ag3Sn is formed by the dissolution of Ag into solder. The crack at the interface of Ag3Sn and solder is easily created by large thermal expansion difference between the Cu ribbon and Si cell. A SEM image of a crack generated at the rear side of the solar cell inside the PbSn solder is shown in Fig. 5.11 [44]. From the SEM image, the grain size of the PbSn solder sandwiched by Ag electrode and Cu ribbon becomes larger than the solder at the back side

Figure 5.10 The cross-sectional view of Cu ribbon interconnection crack under optical microscope. From U. Itoh, M. Yoshida, H. Tokuhisa, K. Takeuchi, Y. Takemura, Solder joint failure modes in the conventional crystalline si module. Energy Procedia 55 (2014) 464e468.

Figure 5.11 The cross-sectional view of solder bond crack observed by SEM [44].

of the Cu ribbon, which is not sandwiched by the Ag electrode. It clearly shows that the Pb and Sn grains initiate growth during the thermal cycles. The large CTE differences between the Cu and solder joint introduce thermal stress during thermal cycles. The solder joint failure is observed visually as microcracks initially. In the PbSn solder, there are two grains of a-Pb and b-Sn. The grain size grows as the number of thermal cycles increase. The bonding strength decreases with increasing thermal cycles. The crack grows at the interface between the large grains.

5.3.2 Characterization Generally, fatigue modeling consists of four primary steps which provide a basis for an otherwise confusing process [48]. First, a theoretical or constitutive equation is defined as the basis for modeling. Appropriate assumptions need to be made in constructing the constitutive equation. Second, the constitutive equation is translated into a finite element analysis (FEA) program (e.g., COMSOL multiphysics software) [49], and a model is created. The FEA program calculates the predicted stresse strain values for the system under study and returns stress values for the simulated conditions. Third, the FEA results are used to create a model predicting the number of cycles to failure (Nf). Fourth, the model or results must be tested and verified using thermal cycling data. These four steps describe the general process for fatigue modeling.

5: D EGRADATION

AND

FAILURE M ECHANISMS

OF

PV M ODULE I NTERCONNECTS

In recent research, N. Bosco et al. have developed an FEM of a flat plate c-Si module to calculate the accumulation of inelastic strain energy density within the solder joint for cyclic temperature exposure [50,51]. These works elucidated the effect of different PV module design and materials as well as the climatic conditions on PbSn eutectic solder joint thermal fatigue durability. The thickness of the solder layer itself has the largest influence on its damage accumulation. If the solder layer is half as thick, the damage accumulated would be approximately twice as much [50]. Increasing the thickness of the copper ribbon and silicon would also increase solder joint damage [50]. Three meteorological factors impact the rate of solder fatigue: mean daily maximum cell temperature, mean daily maximum cell temperature change, and number of temperature reversals across a characteristic temperature 56.4
C. In order to accumulate the equivalent amount of 25 years field exposure damage, in the most damaging city (Chennai, India), it would require 630 accelerated thermal cycles (e40 to 85
C) [51].

5.4 Ribbon Fatigue Interconnect ribbon failure is another common failure mode in flat plate PV modules. Interconnect ribbon fatigue leads to power degradation in terms of increase in the series resistance Rs. Ribbon wire for PV modules is wide and flat shape metal wire flashed copper and solder as shown in Fig. 5.12. The width of

Figure 5.12 Image of the copper ribbon wire for photovoltaic modules. From J. Jeong, N. Park, W. Hong, C. Han, Analysis for the degradation mechanism of photovoltaic ribbon wire under thermal cycling. In: Photovoltaic Specialists Conference (PVSC), 2011 37th IEEE, IEEE, 2011, pp. 003159e003161.

127

the ribbon is constrained to minimize the shadowing loss in a cell. The increase in the width of interconnection ribbon cross section increases the shadowing losses proportionally [52]. The thickness of ribbon strip is limited by buildup stresses in the solder joint. Cracks originated from ribbon fatigue reduce the cross-sectional area of the wire and increase the Rs of the module.

5.4.1 Mechanisms The driving force of this type of failure is known to be metal fatigue, originated from thermally and mechanically induced strain [12,14]. In a typical buildup of a standard crystalline module, as shown in Fig. 5.1, the copper ribbons are soldered to connect the front Ag electrode of a solar cell and the back-side contact of the next solar cell. Through the lamination process, all the cells are encapsulated with the polymeric layer (EVA) sandwiched by glass cover sheet and polymer backsheet. The CTE of glass is almost two times higher than that of c-Si cells. Therefore, a change in temperature generates stress, which leads to a displacement of solar cells in the flexible layer of EVA and loads the copper ribbon in between. At moderate temperatures (above 20
C), EVA is far above the glass transition point. There has a very low modulus of elasticity influence in the intercell displacement. Due to the higher CTE of glass, the cells are pulled apart when temperature rises. As experimentally showed by Meier et al. [54], the growth of the gap is linear to the temperature increase. During cooling, the opposite effect leads to a shrinkage of the gap. In the low-temperature regime (below 20
C), EVA has a much higher modulus of elasticity. The cells are tensioned by polymer during cooling process. Therefore the thermal expansion of the glass has less impact on the PV cells. Although copper ribbon has a higher CTE than glass, the part of ribbon between the two cells is still tensioned by the cell displacement during temperature change, since only a small part of the ribbon can freely expand, the rest is soldered to the cell. At the gap between cells, ribbon is bent during contraction. This bending leads to regions of increased stress, and plastic strain occurs even at low displacement amplitudes. Hardening effects of the copper distribute the stress to the next surrounding and generate a region of hardened material where microcracks are formed in following thermal cycles. Finally, these microcracks combine into large crack and grow in

128

D URABILITY AND R ELIABILITY

OF

P OLYMERS

subsequent thermal cycles until the cross section of the ribbon is reduced to a level that is not able to withstand the applied stress and rupture occurs.

5.4.2 Characterization N. Bosco et al. has experimentally demonstrated that in a better designed PV modules (i.e., copper ribbons soldered to solar cells with some offset) ribbons experience less strain than designed modules (i.e., ribbons soldered to the edge of the cells) [14]. In a more robust design, the ribbon was soldered 10 mm away from the cell edge on the front side and 15 mm away from the cell edge on the back side (Fig. 5.13). Compared to the design that soldered the ribbon right at the edge, the effective gauge length of the interconnect ribbons soldered with offsite is 25 mm longer; therefore, they experience proportionally less strain for a similar amount of cellecell deflection. The experimental result showed that with the ribbons soldered without an offset significant failures occurred roughly 40% faster than the ribbons soldered with an offset.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

5.5 Electrically Conductive Adhesives in Novel Cell Interconnection Strategies Advances in PV module technology have always been pursued in research and development, but many times these new approaches do not thrive in the market because of the challenges of lab-based qualification and real-world lifetime performance [55]. Two cell technology approaches, which both aimed to replace ribbon-based cell interconnection, are back-contact cells such as metal or emitter wrapthrough (WT) and more recently shingled full or sliced (half or “sliver”) cells [56]. Both of these approaches have the goal of incorporating a new material class, ECAs, into PV modules [57]. ECAs offer the promise of reduced module assembly costs, which is an ongoing industry goal, but bringing new technologies to commercial scale with demonstrated reliability is a large challenge. Efforts from 2000 to 2010 focused on commercializing WT cells in PV modules were unsuccessful, due to cost and reliability issues [58]. The current focus on shingled cells appears to be progressing well into commercial production.

5.5.1 Wrap-Through Cells and Their Interconnection

Figure 5.13 Illustration of the solder point at interconnect ribbon on each cell with and without an offset. From N. Bosco, T.J. Silverman, J. Wohlgemuth, S. Kurtz, M. Inoue, K. Sakurai, et al., Evaluation of dynamic mechanical loading as an accelerated test method for ribbon fatigue. In: Photovoltaic Specialists Conference (PVSC), 2013 IEEE 39th, IEEE, 2013, pp. 3173e3178.

Back-contact cells remove the need for front-side cell metallization and their associated busbars but introduce more complexity in having to accommodate the front and backside interconnections on the cell backside [59]. The fabrication and manufacturing advantages were considered positively in the early 2000s, with the idea that pick and place automated assembly methods used in the consumer electronics industry would simplify module manufacturing [58,60]. This approach relies on a polymeric cell “carrier sheet” with printed or electroplated metallization, analogous to a printed wiring or circuit board. The cells are then placed onto ECA droplets on the wiring spots and allowed to cure. There is a distinct set of challenges in shifting the industry from metallic solder-based cell connection and replacing it with polymeric adhesive cell connections, for example, the qualification and

5: D EGRADATION

AND

FAILURE M ECHANISMS

OF

PV M ODULE I NTERCONNECTS

reliability standards are all focused on the metallic solder failure mechanisms, and the current test conditions are be wholly inappropriate for an ECA [61]. In addition, such dramatic changes in not only materials but also interconnect geometries can lead to undesirable changes from familiar electrical characteristics [62]. These back-contact cell technologies appear to have been broadly rejected by high-volume manufacturers at this time, possibly due to high costs and the qualification and reliability issues that arise.

5.5.2 Shingled Cells and Subcells Interconnected With ECAs More recently, shingling of cells in the PV cells is being pursued as a new technology option for manufacturers [63]. This follows the industry-wide shift from Al-BSF cells to PERC cells, in which both of these cells utilize the same front- and back-side metallization as has been commercial for 30 years (refer Chapter 4). Shingling provides advantages of increasing voltages, decreasing current, and the associated I2R losses in a module. Shingling is also compatible with bifacial cells [64], such as the bifacial PERC cells that have recently been commercialized [65,66]. And in these shingled cell modules, ECAs play an important role in enabling a new interconnection architecture [67].

5.5.3 Performance of PV Modules Utilizing ECAs The performance of PV modules using ECAs has been successful enough to motivate ongoing research and development, for example, during the period of research focusing on the back-contact cell approaches [68]. And in the case of shingled cells, initial results for performance and reliability have been positive and motivated increased research efforts [69].

5.5.4 Qualification and Reliability of ECA Materials and PV Modules With the positive performance of shingled cell modules using ECAs, there critical issues of reliability and lifetime performance have become the focus of research. For example, the necessary material properties of ECAs for shingled cell PV modules

129

are critical as are electrical and mechanical criteria [70]. At the same time, qualification tests for PV module ECAs do not exist. There is still much work necessary on developing the database of ECA materials testing results to define the necessary qualification tests, guide materials selection, inform manufacturing processes [71]. It is advantageous that many materials systems, such as silicones, have a long and broad history of use, and that ECAs were studied for the back-contact cells in the early 2000s. This historic information rapidly informs the current ECAs for shingled cell modules [57]. In addition, it is essential to develop models and methods to optimize ECA selection and development for shingle cell modules, and this work is proceeding apace [72]. And the PV research community is now starting to establish the foundation of reported ECA failures in module tests, identify both root causes, and characterization methods to address these [73]. For example, electrical studies inform back-contact cells current studies [74]. And similarly, mechanical tests and failures can help inform the PV community [75,76]. Novel interconnect approaches, such as ECAs for shingled cell PV modules, look today to be promising and are actively being researched, but their true role in our terawatt PV power system has yet to be established.

5.6 Conclusions This chapter reviewed the major degradation and failures of interconnections in silicon PV modules, which include solder joint thermomechanical fatigue and interconnect ribbon fatigue. Ag leaching into solder and long-term thermal fatigue are identified as two major causes of solder failures. Fatigue of metal, induced by mechanical loads on ribbon wire due to thermal cycles, leads to interconnect ribbon failures. In addition, the recent interest in novel interconnect strategies, such as the use of ECAs to replace the tinned copper ribbons, opens up large new areas of both module cost and performance. These ECAs represent a new manifold of degradation mechanisms for these new interconnection materials and new module architectures, for which we do not have sufficient field or lab experience to make predictions of 30 year lifetime performance.

130

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Acknowledgments The authors acknowledge the SETO Office, our academic and industrial collaborators, and all the students.

References [1] T. Saga, Advances in crystalline silicon solar cell technology for industrial mass production, NPG Asia Materials 2 (3) (2010) 96. [2] M.T. Zarmai, N. Ekere, C. Oduoza, E.H. Amalu, A review of interconnection technologies for improved crystalline silicon solar cell photovoltaic module assembly, Applied Energy 154 (2015) 173e182. [3] R. Klengel, M. Petzold, D. Schade, B. Sykes, Improved testing of soldered busbar interconnects on silicon solar cells, in: Microelectronics and Packaging Conference (EMPC), 2011 18th European, IEEE, 2011, pp. 1e5. [4] C.E. Packard, J.H. Wohlgemuth, S.R. Kurtz, Development of a Visual Inspection Data Collection Tool for Evaluation of Fielded PV Module Condition, tech. rep., National Renewable Energy Lab.(NREL), Golden, CO (United States), 2012. [5] P.J. Gress, P.I. Widenborg, S. Varlamov, A.G. Aberle, Wire bonding as a cell interconnection technique for polycrystalline silicon thin-film solar cells on glass, Progress in Photovoltaics: Research and Applications 18 (3) (2010) 221e228. [6] S. Sakamoto, T. Kobayashi, S. Nonomura, Epidemiological analysis of degradation in silicon photovoltaic modules, Japanese Journal of Applied Physics 51 (10S) (2012) 10NF03. [7] H. Xiong, C. Gan, X. Yang, Z. Hu, H. Niu, J. Li, et al., Corrosion behavior of crystalline silicon solar cells, Microelectronics Reliability 70 (Supplement C) (2017) 49e58, https://doi.org/10.1016/ j.microrel.2017.01.006. http://www.sciencedirect. com/science/article/pii/S0026271417300069. [8] C. Peike, S. Hoffmann, P. Hlsmann, K.A. Wei, M. Koehl, P. Bentz, Permeation impact on metallization degradation, in: N.G. Dhere, J.H. Wohlgemuth (Eds.), SPIE Solar Energy þ Technology, 2012, pp. 84720Ue84729U, https://doi.org/10.1117/12.929819, in: http:// proceedings.spiedigitallibrary.org/proceeding. aspx?articleid¼1379672.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

[9] M.C.C. Oliveira, A.S.A. Cardoso Diniz, M.M. Viana, V.F.C. Lins, The causes and effects of degradation of encapsulant ethylene vinyl acetate copolymer (EVA) in crystalline silicon photovoltaic modules: a review, Renewable and Sustainable Energy Reviews 81 (P2) (2017) 2299e2317, https://doi.org/10.1016/j.rser.2017. 06.039. http://www.sciencedirect.com/science/ article/pii/S1364032117309851. [10] J.S. Jeong, N. Park, C. Han, Field failure mechanism study of solder interconnection for crystalline silicon photovoltaic module, Microelectronics Reliability 52 (9e10) (2012) 2326e2330. [11] F.P. McCluskey, Reliability modeling for photovoltaic modules, in: NREL Photovoltaic Modules Reliability Workshop, Colorado, 2010, pp. 75e97. https://www.nrel.gov/docs/fy14osti/ 60171.pdf. [12] L.N. Dumas, A. Shumka, Photovoltaic module reliability improvement through application testing and failure analysis, IEEE Transactions on Reliability 31 (3) (1982) 228e234. [13] M. Smokler, D. Otth, R. Ross Jr., The block program approach to photovoltaic module development, in: Proceedings of the 18th IEEE PV Specialists Conference, 1985, pp. 1150e1158. [14] N. Bosco, T.J. Silverman, J. Wohlgemuth, S. Kurtz, M. Inoue, K. Sakurai, et al., Evaluation of dynamic mechanical loading as an accelerated test method for ribbon fatigue, in: Photovoltaic Specialists Conference (PVSC), 2013 IEEE 39th, IEEE, 2013, pp. 3173e3178. [15] D.L. Meier, V. Chandrasekaran, A. Gupta, V. Yelundur, A. Rohatgi, Silver contact grid: inferred contact resistivity and cost minimization in 19% silicon solar cells, IEEE Journal of Photovoltaics 3 (1) (2013) 199e205. [16] J.S. Yu, I. Kim, J.S. Kim, J. Jo, T.T. LarsenOlsen, R.R. Søndergaard, et al., Silver front electrode grids for ito-free all printed polymer solar cells with embedded and raised topographies, prepared by thermal imprint, flexographic and inkjet roll-to-roll processes, Nanoscale 4 (19) (2012) 6032e6040. [17] T. Tanahashi, N. Sakamoto, H. Shibata, A. Masuda, Electrical detection of gap formation underneath finger electrodes on c-Si PV cells exposed to acetic acid vapor under hygrothermal conditions, in: 2016 IEEE 43rd Photovoltaic

5: D EGRADATION

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

AND

FAILURE M ECHANISMS

OF

PV M ODULE I NTERCONNECTS

Specialists Conference, PVSC), 2016, pp. 1075e1079. https://doi.org/10.1109/PVSC. 2016.7749778. M. Kntges, G. Oreski, U. Jahn, M. Herz, P. Hacke, K.-A. Weiss, et al. Assessment of Photovoltaic Module Failures in the Field. Tech. Rep. T13-092017, IEA-PVPS Task 13, 2017. URL: http:// www.iea-pvps.org/index.php?id¼435. M. Ko¨ntges, S. Kurtz, C. Packard, U. Jahn, K. Berger, K. Kato, et al., Review of failures of photovoltaic modules, IEA PVPS Task 13 (2014). C. Peike, S. Hoffmann, P. Hlsmann, B. Thaidigsmann, K.A. Wei, M. Koehl, et al., Origin of damp-heat induced cell degradation, Solar Energy Materials and Solar Cells 116 (Supplement C) (2013) 49e54. https://doi.org/ 10.1016/j.solmat.2013.03.022. http://www. sciencedirect.com/science/article/pii/S0927024 813001359. IEC, IEC 61215 Terrestrial Photovoltaic (PV) Modules e Design Qualification and Type Approval. International Standard, International Electrotechnical Commission, 2016. https:// webstore.iec.ch/publication/24312. J.H. Wohlgemuth, S. Kurtz, Using accelerated testing to predict module reliability, in: Photovoltaic Specialists Conference (PVSC), 2011 37th IEEE, IEEE, 2011, pp. 003601e003605. P. Hacke, K. Terwilliger, S. Glick, D. Trudell, N. Bosco, S. Johnston, et al., Test-to-Failure of crystalline silicon modules, in: 2010 35th IEEE Photovoltaic Specialists Conference, 2010, pp. 000244e000250. https://doi.org/10.1109/ PVSC.2010.5614472. M.A. Quintana, D.L. King, T.J. McMahon, C.R. Osterwald, Commonly observed degradation in field-aged photovoltaic modules, in: Conference Record of the Twenty-Ninth IEEE Photovoltaic Specialists Conference, 2002, 2002, pp. 1436e1439. https://doi.org/10.1109/ PVSC.2002.1190879. W. Zhang, P.B. Unit, How Silver Paste Can Improve Silicon Solar Cell Performance/cost Ratio, Hereaus Materials Technology LLT, 2011. M.D. Kempe, G.J. Jorgensen, K.M. Terwilliger, T.J. McMahon, C.E. Kennedy, T.T. Borek, Acetic acid production and glass transition concerns with ethylene-vinyl acetate used in photovoltaic

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

131

devices, Solar Energy Materials and Solar Cells 91 (4) (2007) 315e329. A. Kraft, L. Labusch, T. Ensslen, I. Du¨rr, J. Bartsch, M. Glatthaar, et al., Investigation of acetic acid corrosion impact on printed solar cell contacts, IEEE Journal of Photovoltaics 5 (3) (2015) 736e743. D. Chenvidhya, K. Kirtikara, C. Jivacate, Pv module dynamic impedance and its voltage and frequency dependencies, Solar Energy Materials and Solar Cells 86 (2) (2005) 243e251. T. Tanahashi, N. Sakamoto, H. Shibata, A. Masuda, Localization and characterization of a degraded site in crystalline silicon photovoltaic cells exposed to acetic acid vapor, IEEE Journal of Photovoltaics 8 (4) (2018) 997e1004. https:// doi.org/10.1109/JPHOTOV.2018.2839259. C. Ballif, D. Huljic, G. Willeke, A. HesslerWyser, Silver thick-film contacts on highly doped n-type silicon emitters: structural and electronic properties of the interface, Applied Physics Letters 82 (12) (2003) 1878e1880. Z. Li, L. Liang, A. Ionkin, B. Fish, M. Lewittes, L. Cheng, et al., Microstructural comparison of silicon solar cells front-side ag contact and the evolution of current conduction mechanisms, Journal of Applied Physics 110 (7) (2011) 074304. H. Morita, M. Miyashita, A. Masuda, Moisture ingress rate and route into c-si pv modules, in: Photovoltaic Module Reliability Workshop 2014, 2014, p. 601. A. Masuda, N. Uchiyama, Y. Hara, Degradation by acetic acid for crystalline si photovoltaic modules, Japanese Journal of Applied Physics 54 (4S) (2015) 04DR04. N. Park, W. Oh, D. Kim, Effect of temperature and humidity on the degradation rate of multicrystalline silicon photovoltaic module, International Journal of Photoenergy 2013 (2013). M. Koehl, M. Heck, S. Wiesmeier, Modelling of conditions for accelerated lifetime testing of humidity impact on pv-modules based on monitoring of climatic data, Solar Energy Materials and Solar Cells 99 (2012) 282e291. M.D. Kempe, J.H. Wohlgemuth, Evaluation of temperature and humidity on pv module component degradation, in: Photovoltaic Specialists Conference (PVSC), 2013 IEEE 39th, IEEE, 2013, pp. 0120e0125.

132

D URABILITY AND R ELIABILITY

OF

P OLYMERS

[37] P. Hu¨lsmann, K.A. Weiss, Simulation of water ingress into pv-modules: Iec-testing versus outdoor exposure, Solar Energy 115 (2015) 347e353. [38] S. Suzuki, T. Tanahashi, T. Doi, A. Masuda, Acceleration of degradation by highly accelerated stress test and air-included highly accelerated stress test in crystalline silicon photovoltaic modules, Japanese Journal of Applied Physics 55 (2) (2016) 022302. [39] Y. Voronko, G.C. Eder, M. Knausz, G. Oreski, T. Koch, K.A. Berger, Correlation of the loss in photovoltaic module performance with the ageing behaviour of the backsheets used, Progress in Photovoltaics: Research and Applications 23 (11) (2015) 1501e1515. [40] G. Oreski, A. Rauschenbach, C. Hirschl, M. Kraft, G. Eder, G. Pinter, Crosslinking and post-crosslinking of ethylene vinyl acetate in photovoltaic modules, Journal of Applied Polymer Science 134 (23) (2017). [41] K. Whitfield, A. Salomon, S. Yang, I. Suez, Damp heat versus field reliability for crystalline silicon, in: Photovoltaic Specialists Conference (PVSC), 2012 38th IEEE, IEEE, 2012, pp. 001864e001870. [42] M. Ko¨ntges, G. Oreski, U. Jahn, M. Herz, P. Hacke, K. Weiss, et al., Assessment of Photovoltaic Module Failures in the Field, International Energy Agency, Paris, France, 2017. [43] D. King, M. Quintana, J. Kratochvil, D. Ellibee, B. Hansen, Photovoltaic module performance and durability following long-term field exposure, Progress in Photovoltaics: Research and Applications 8 (2) (2000) 241e256. [44] U. Itoh, M. Yoshida, H. Tokuhisa, K. Takeuchi, Y. Takemura, Solder joint failure modes in the conventional crystalline si module, Energy Procedia 55 (2014) 464e468. [45] W. Bader, Dissolution of au, ag, pd, pt., cu and ni in a molten tin-lead solder, Welding Research Supplement (1969) 551e557. [46] T. Yang, K. Huang, S. Yang, H. Hsieh, C. Kao, Growth kinetics of ag3sn in silicon solar cells with a sintered ag metallization layer, Solar Energy Materials and Solar Cells 123 (2014) 139e143.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

[47] G. Cuddalorepatta, A. Dasgupta, S. Sealing, J. Moyer, T. Tolliver, J. Loman, Durability of pb-free solder between copper interconnect and silicon in photovoltaic cells, Progress in Photovoltaics: Research and Applications 18 (3) (2010) 168e182. [48] W. Lee, L. Nguyen, G. Selvaduray, Solder joint fatigue models: review and applicability to chip scale packages, Microelectronics Reliability 40 (2) (2000) 231e244. https://doi.org/10.1016/ S0026-2714(99)00061-X. http://www.science direct.com/science/article/pii/S002627149900 061X. [49] C. Multiphysics, Comsol multiphysics user guide (version 4.3 a), COMSOL, AB (2012) 39e40. [50] N. Bosco, T.J. Silverman, S.R. Kurtz, The influence of pv module materials and design on solder joint thermal fatigue durability, IEEE Journal of Photovoltaics 6 (2016) 1407e1412. [51] N. Bosco, T.J. Silverman, S. Kurtz, Climate specific thermomechanical fatigue of flat plate photovoltaic module solder joints, Microelectronics Reliability 62 (2016) 124e129. [52] E.V. Kerschaver, G. Beaucarne, Back-contact solar cells: a review, Progress in Photovoltaics: Research and Applications 14 (2) (2006) 107e123. [53] J. Jeong, N. Park, W. Hong, C. Han, Analysis for the degradation mechanism of photovoltaic ribbon wire under thermal cycling, in: Photovoltaic Specialists Conference (PVSC), 2011 37th IEEE, IEEE, 2011, pp. 003159e003161. [54] R. Meier, F. Kraemer, S. Wiese, K.J. Wolter, J. Bagdahn, Reliability of copper-ribbons in photovoltaic modules under thermo-mechanical loading, in: Photovoltaic Specialists Conference (PVSC), 2010 35th IEEE, IEEE, 2010, pp. 001283e001288. [55] M.H. Alaaeddin, S.M. Sapuan, M.Y.M. Zuhri, E.S. Zainudin, F.M. AL-Oqla, Photovoltaic applications: Status and manufacturing prospects, Renewable and Sustainable Energy Reviews 102 (2019) 318e332. https://doi.org/10.1016/j.rser. 2018.12.026. http://www.sciencedirect.com/ science/article/pii/S1364032118308219. [56] M.T. Zarmai, N.N. Ekere, C.F. Oduoza, E.H. Amalu, A review of interconnection

5: D EGRADATION

[57]

[58]

[59]

[60]

[61]

[62]

[63]

AND

FAILURE M ECHANISMS

OF

PV M ODULE I NTERCONNECTS

technologies for improved crystalline silicon solar cell photovoltaic module assembly, Applied Energy 154 (2015) 173e182. https:// doi.org/10.1016/j.apenergy.2015.04.120. http:// www.sciencedirect.com/science/article/pii/S03 06261915005838. G. Beaucarne, K. Broek, B. Chislea, Y. Yu, J. Wei, A. Zambova, et al., Study of compatibility of silicone-based electrically conductive adhesives and conductive backsheets for MWT modules, Energy Procedia 55 (2014) 444e450. https://doi.org/10.1016/j.egypro.2014.08.007. P.C. de Jong, Achievements and challenges in crystalline silicon back-contact module technology, in: Photovoltaics International, seventh ed., 2010, pp. 138e144. A. Schonecker, D. Eikelboom, P. Manshanden, M. Goris, P. Wyers, S. Roberts, et al., ACE Designs: the beauty of rear contact solar cells, in: Conference Record of the Twenty-Ninth IEEE Photovoltaic Specialists Conference, 2002, 2002, pp. 106e109. https://doi.org/10.1109/ PVSC.2002.1190467. J.H. Bultman, D.W.K. Eikelboom, R. Kinderman, A.C. Tip, C.J.J. Tool, MACJvd Nieuwenhof, et al., Fast and easy single step module assembly for back-contacted c-Si solar cells with conductive adhesives, in: Proceedings of 3rd World Conference on Photovoltaic Energy Conversion, 2003, vol. 1, 2003, pp. 979e982, vol. 1. R. Witteck, H. Schulte-Huxel, H. Holst, D. Hinken, M. Vogt, S. Blankemeyer, et al., Testing interconnections using conductive adhesives for application in PV modules, in: 21st European Photovoltaic Solar Energy Conference, Dresden, Germany, 2006, pp. 2137e2142. A. Fallisch, D. Biro, 2-D SPICE simulation and analytical calculation of spreading resistance effects in emitter wrap-through cells with nonsquare via-hole pattern, IEEE Journal of Photovoltaics 1 (2) (2011) 153e158. https://doi. org/10.1109/JPHOTOV.2011.2172190. D. Tonini, G. Cellere, M. Bertazzo, A. Fecchio, L. Cerasti, M. Galiazzo, Shingling technology for cell interconnection: technological aspects and process integration, Energy Procedia 150 (2018) 36e43. https://doi.org/10.1016/j.egypro. 2018.09.010. http://www.sciencedirect.com/ science/article/pii/S1876610218305502.

133

[64] N. Klasen, A. Mondon, A. Kraft, U. Eitner, Shingled Cell Interconnection: A New Generation of Bifacial PV-Modules, SSRN Scholarly Paper ID 3152478, Social Science Research Network, Rochester, NY, 2018, https://papers. ssrn.com/abstract¼3152478. [65] J.L. Braid, T.D. Wager, A. Longacre, B.D. Huey, R.H. French, Initial stability of PERC vs. Al-BSF cells, in: 2018 IEEE 7th World Conference on Photovoltaic Energy Conversion (WCPEC) (A Joint Conference of 45th IEEE PVSC, 28th PVSEC 34th EU PVSEC), 2018, pp. 1255e1260. https://doi.org/10.1109/PVSC. 2018.8547401. [66] J.L. Braid, A.J. Curran, J. Sun, E.J. Schneller, J.S. Fada, J. Liu, et al., EL and I-V Correlation for degradation of PERC vs. Al-BSF commercial modules in accelerated exposures, in: 2018 IEEE 7th World Conference on Photovoltaic Energy Conversion (WCPEC) (A Joint Conference of 45th IEEE PVSC, 28th PVSEC 34th EU PVSEC), 2018, pp. 1261e1266. https://doi.org/ 10.1109/PVSC.2018.8547420. [67] L. Theunissen, B. Willems, J. Burke, D. Tonini, M. Galiazzo, A. Henckens, Electrically conductive adhesives as cell interconnection material in shingled module technology, AIP Conference Proceedings 1999 (1) (2018) 080003. https://doi. org/10.1063/1.5049305. https://aip.scitation.org/ doi/abs/10.1063/1.5049305. [68] M. Meuwissen, M. Van Den Nieuwenhof, H. Steijvers, T. Bots, K. Broek, M. Kloos, Simulation assisted design of a PV module incorporating electrically conductive adhesive interconnects, in: 21st European Photovoltaic Solar Energy Conference and Exhibition, Dresden, Germany, 2006, pp. 2485e2490. [69] H. Hsieh, W. Lee, T. Chang, C. Hsi, Photovoltaic modules of crystalline solar cells using a new type assembly structure, in: 2010 5th International Microsystems Packaging Assembly and Circuits Technology Conference, 2010, pp. 1e4. https://doi.org/10.1109/IMPACT.2010.5699598. [70] G. Beaucarne, Materials challenge for shingled cells interconnection, Energy Procedia 98 (2016) 115e124. https://doi.org/10.1016/j. egypro.2016.10.087. http://www.sciencedirect. com/science/article/pii/S1876610216310487. [71] L.P. Bauermann, E. Fokuhl, S. Stecklum, D. Philipp, T. Geipel, A. Kraft, et al.,

134

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Qualification of conductive adhesives for photovoltaic application e accelerated ageing tests, Energy Procedia 124 (2017) 554e559. https://doi.org/10.1016/j.egypro.2017.09.266. http://www.sciencedirect.com/science/article/ pii/S1876610217341978. [72] R. Witteck, D. Hinken, H. Schulte-Huxel, M.R. Vogt, J. Muller, S. Blankemeyer, et al., Optimized interconnection of passivated emitter and rear cells by experimentally verified modeling, IEEE Journal of Photovoltaics 6 (2) (2016) 432e439. https://doi.org/10.1109/ JPHOTOV.2016.2514706. http://ieeexplore. ieee.org/document/7390201/. [73] J. Wang, Z. Guo, H. Jin, Eca (electrical conductive adhesive) induced failure on shingling module, in: Light, Energy and the Environment 2018 (E2, FTS, HISE, SOLAR, SSL) (2018), Paper JT2A.10, Optical Society of America, 2018, p. JT2A.10. https://doi.org/10.1364/FTS. 2018.JT2A.10. https://www.osapublishing.org/ abstract.cfm?uri¼HISE-2018-JT2A.10.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

[74] R. Witteck, H. Schulte-Huxel, H. Holst, D. Hinken, M. Vogt, S. Blankemeyer, et al., Testing interconnections using conductive adhesives for application in PV modules, in: 21st European Photovoltaic Solar Energy Conference, Dresden, Germany, 2006, pp. 2137e2142. [75] S. Wiese, F. Kraemer, N. Betzl, D. Wald, Interconnection technologies for photovoltaic modules-analysis of technological and mechanical problems, in: Thermal, Mechanical & Multi-Physics Simulation, and Experiments in Microelectronics and Microsystems (EuroSimE), 2010 11th International Conference on, IEEE, 2010, pp. 1e6. http://ieeexplore.ieee.org/ abstract/document/5464518/. [76] S. Hoffmann, T. Geipel, M. Meinert, A. Kraft, Analysis of peel and shear forces after temperature cycle tests for electrical conductive adhesives, in: Proceedings of the 33rd European Photovoltaic Solar Energy Conference and Exhibition, 2017, pp. 183e186.

6 Degradation Processes and Mechanisms of Encapsulants Gernot Oreski, Bettina Ottersbo¨ck and Antonia Omazic Polymer Competence Center Leoben, Leoben, Austria

6.1 State of the Art: PV Module Encapsulation Using Ethylene Vinyl Acetate (EVA) The general module constitution for crystalline silicon PV modules has been developed in the FlatPlate Solar Array Project in the 1970s and 1980s and has not significantly changed since then [1]. The primary purpose of the encapsulant is to bond the multiple components of a module together. Especially the surrounding of the brittle solar cells with its electrical connections is very important to give also the required protection. Within a PV module the encapsulation material has to fulfill several basic functions. These include providing structural support and physical isolation of the solar cells, maintaining electrical isolation, and being highly transparent in a selected spectral region, according to the cell technology used. To deal with different thermal expansions of the materials used in a module (glass, solar cell, and interconnects) and to avoid overstressing and cracking, the material has to be a low modulus, elastomeric material. Furthermore, for PV modules a lifetime of 20e30 years with a maximum total loss in module performance of 20% over a 20 year period needs to be guaranteed. Table 6.1 summarizes the requirements for solar cell encapsulation materials. Since the late 1980s the dominating encapsulation material for PV modules is the chemically crosslinked ethylene vinyl acetate copolymer (EVA). Compared to previous solutions based on silicone rubbers (polydimethyl siloxanedPDMS) and polyvinyl butyral (PVB) [3], EVA films have some specific advantages like easy handling, good optical and mechanical properties, low weight, and low-cost production [2]. EVA is a semicrystalline random copolymer of ethylene and vinyl acetate (VA). When used as a solar

cell encapsulation film, the vinyl acetate content ranges usually from 28% to 33%. The melting temperature for PV grade EVA types is between 60 and 70
C, and the lamination temperature is around 150
C, depending on the peroxide curing agent which initiates the cross-linking reaction. During lamination a three dimensional network is formed by chemical cross-linking of the polymer chains in order to increase the thermal stability of the material and to prevent the material from melting within application relevant temperatures up to 100
C [2,4]. By crosslinking the copolymer chains (VA) during PV module lamination, the EVA sheet is transformed into an elastomeric, transparent, and dimensionally stable encapsulation material. The curing process is initiated via a radical reaction, using an organic peroxide or peroxycarboxylic acid as a radical initiator (“crosslinker” or “curing agent”) [2,4]. The initial degree of crosslinking after lamination is very important as it strongly influences the longterm characteristics of the material and is therefore highly relevant for quality and long term stability of the produced PV module [2,5]. It is controlled by the lamination temperature, which affects the amount of crosslinker activated per time unit, the lamination time, and the initial concentration. After crosslinking EVA has a melting range between 30 and 70
C with calculated crystallinities of approximately 7%e9% [4]. However, due to the chemical crosslinks the melting of the crystallites is not followed by a softening or creep of the encapsulation film. The crosslinking of EVA also results in an increased solar transmittance to values between 88% and 90%, depending on the type of the UV absorber used. A recent development is the introduction of UV absorbers with a lower UV cut-off (20 years old modules are most likely for high quality modules. Therefore, this group of modules shows the lowest degradation rate. This effect can be seen in Fig. 8.68, which shows the degradation rates from Fig. 8.67 partitioned by the length of field exposure. For the modules with an exposure time of up to 10 years, the rate distribution has a much more pronounced tail and a higher median than the modules exposed for more than 10 years. Chicca et al. compared the M55 crystalline silicon modules from Arco exposed in California (temperate climate) for 28 years and the same type of modules from Siemens exposed in Arizona (Hot & Dry climate) for 18 years. They found that the average power degradation rate for 18 years old module was 1.17%/year while it was 0.39% for the 28 years old module [52]. This study shows that the longer installation years do not necessarily lead to higher degradation rates. Theoretically, to understand

211

Figure 8.68 Degradation rate histogram grouped by outdoor exposure length. The median rate for the exposure length up to 10 years is significantly higher than that for the length of 10 years and longer. Fig. 3 of Ref. [17].

the impact of exposure time to the power degradation rate, one has to analyze the same modules with the same location without maintenance (module replacement) since different environmental stresses will have different impacts on the module performance. The section below will discuss the impact of the climate zones.

8.3.4 Module Power Degradation Rate by Climate Zones As mentioned before, module power degradation is environmental stress dependent. Christopher Raupp et al. [47] studied the degradation rates of approximately 59,000 PV modules from 26 operational PV power plants in various climatological regions of the United States (ArizonadHot-dry; CaliforniadTemperate; ColoradodTemperate; New YorkdCold-dry; and TexasdHot-humid). The systems range from about 1 year to 19 years in age. The evaluated PV power plants represent about 252,000 PV modules of the following commercially available technologies: mono-Si, poly-Si, hetero junction silicon (HIT), amorphous silicon (a-Si), cadmium telluride (CdTe), and copper indium gallium selenide (CIGS). The average annual degradation rates of c-Si modules in hot-dry and hot-humid climates are higher than 1%/year and in temperate and cold-dry climates are lower than 0.7%/year, as

212

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 8.69 Degradation rates of IeV parameters of c-Si modules in all climates. Redrawn Fig. 4 of C. Raupp, C. Libby, S. Tatapudi, D. Srinivasan, J. Kuitche, B. Bicer, G. Tamizhmani, Degradation Rate Evaluation of Multiple PV Technologies from 59,000 Modules Representing 252,000 Modules in Four Climatic Regions of the United States, IEEE, 2016.

highest degradation rates of 1.29% and 1.55% respectively, while the cold/dry zone has the lowest rate of 0.19%. The hot/dry zone shows the highest degradation rate because the high temperatures of the desert-type climate lead to increased EVA browning which manifests itself in high Isc degradation [6]. The highest rate is also caused by the temperature cycling effect due to the large day/night temperature range in this climatic zone. The temperature cycling effect empirically follows the Coffin-Manson equation: N ¼ d/(Dtemp) b1 [54], where N is the number of cycles needed for the material to reach failures. It is clear that the larger the temperature range, the smaller the number of cycles needed to cause failures. Ideally the same type of modules should be compared to illustrate the climatic zone effects. Chicca et al. studied the same M55 crystalline silicon modules of Arco installed in California (temperate climate) for 28 years and in Arizona (hot-dry climate) for 18 years. As expected, they found that the average power degradation rate for Arizona (hotdry) was 1.17%/year while those for the California (temperate climate) was 0.39%/year [52], as shown Ï

shown in Fig. 8.69. This observation is consistent with the environment stress condition because almost all degradations are thermal and moisture activated: the higher the temperature, the higher the relative humidity, the higher the degradation. Kimball et al. analyzed the time-to-failure data from PV modules subjected to damp heat tests under different temperature and relative humidity conditions and found that the time to failure followed the power law model TTF ¼ A$e (Ea/kT)$RH (n), where TTF is the time in hours to reach 20% module power loss, A ¼ 6.4e10, Ea ¼ 0.89  0.11 eV, and n ¼ 2.2  0.8. The test durations in hours at 85
C/85% RH that is expected to correspond to 25 years of operation in Europe, China, and India are about 1000, 500, and 2000 h respectively [53]. This is also the reason that the current IEC standard committee proposes different testing requirements for the different application fields (BC099). Similar degradation behavior was also observed by Rajiv Dubey et al. in the 2013 Indian module survey. Table 8.5 shows the annual power degradation rates for the five different climate zones in their study: the hot/humid and hot/dry zones show the Ï

Ï

8: C RYSTALLINE S ILICON PV M ODULE F IELD FAILURES

213

Table 8.5 The Median Values of the Degradation of the I-V Parameters in the Five Climatic Zones [6] Climatic Zone

Pmax (%/Year)

Isc (%/Year)

Voc (%/Year)

FF (%/Year)

Hot & Humid

1.29

0.98

0.34

0.89

Temperate

0.24

0.34

0.04

0.46

Composite

0.56

0.45

0.17

0.64

Hot & Dry

1.55

0.78

0.20

0.51

Cold & Dry

0.19

0.72

0.11

0.21

Figure 8.70 Comparison of degradation of performance parameters for the same M55 Arco modules installed in Arizona for 18 years and California for 28 years, Fig. 4 of Ref. [52].

in Fig. 8.70, even though the modules in California were exposed 10 years longer than the ones in Arizona. This study definitely demonstrates that module performance is highly climate dependent. The IEA PVPS Task 13 released the report “Assessment of photovoltaic module failures in the field” in 2017 [8], after the study in 2014. They surveyed the industry and collected failure data of PV systems in different climate zones. The power degradation rate caused by a specific module failure mode for a specific climate zone was calculated by converting the measured power loss into a degradation rate. Table 8.6 shows the leading failure modes

in different climate regions and the mean annual power degradation rates caused by them. The most impactful failures on the performance of PV modules for Hot & Humid, Hot & Dry, Moderate, and Cold & Snow are: PID/disconnected cells and strings, defective diodes/burn marks, defective diodes/ disconnected cells and strings, and cell cracks/glass breakage. In summary, this chapter summarizes the major observed field failures of every key BOM components for the typical crystalline silicon PV modules. The most common failures are: junction box failures, glass breakage, defective cell (cell cracks, snail trails, and burn marks) and string interconnect, delamination, loose frame breakage, EVA discoloration, potential induced degradation, and defective bypass diodes. The power degradation of the historical modules is analyzed and is found that about one third of the modules in the field could not meet the regular module warranty terms. The top failures impacting the power degradation are: PID, failure of bypass diodes, cell cracks, and discoloration of the encapsulant and soiling of PV modules. The impact of climate zones on module power degradation is also discussed. It is found that the hot/dry and hot/humid environments have a much higher negative impact on the module power performance compared to the moderate and cold zones. The purpose of this chapter is to provide a summary of PV module failures in the fields and their impacts on power generation, for the PV industry to develop better testing protocols and manufacturing practices to produce more durable PV modules that can achieve 30þ years of field service.

214

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Table 8.6 The Power Degradation Rates of the Leading Failure Modes for the Different Climate Zones, Information Extracted From Fig. 50 of Ref. [8]

delamination

defective bypass diode

burn marks

PID

defect backsheet

defective bypass diode

disconnected cell=string

PID

defective junction box

cell cracks

glass breakage

disconnected cell=string

Encapsulant discoloring

Cold & Snow

Encapsulant discoloring

Moderate

disconnected cell=string

Hot & Dry

PID

Hot & Humid

6.3

2.2

2.0

1.8

11.0

7.7

7.6

1.7

25.0

19.8

15.8

13.4

7.9

6.8

3.1

1.4

References [1] U.S.Department Energy, The History of solar. Available at: https://www1.eere.energy.gov/ solar/pdfs/solar_timeline.pdf. [2] E. Wesoff, IEA: Global Installed PV Capacity Leaps to 303 Gigawatts, 2017. Available at: https://www.greentechmedia.com/articles/read/ iea-global-installed-pv-capacity-leaps-to-303gw#gs.E7rAO0g. [3] Fraunhofer_ISE, Global Market-share in Terms of Annual Production by PV Technology since 1990, 2014. Available at: https://en.wikipedia. org/wiki/Crystalline_silicon#/media/File: Global_Market_Share_by_PV_Technology_ from_1990_to_2013.svg. [4] R. Margolis, D. Feldman, D. Boff, Q4 2016/Q1 2017 Solar Industry Update, 2017. Available at: https://www.nrel.gov/docs/fy17osti/68425.pdf. [5] IEA-PVPS, PV Failure Survey, 2015. Available at: http://www.iea-pvps.org/index.php?id¼9& eID¼dam_frontend_push&docID¼2918. [6] R. Dubey, S. Chattopadhyay, V. Kuthanazhi, J.J. John, B.M. Arora, A. Kottantharayil, K.L. Narasimhan, C.S. Solanki, V. Kuber, J. Vasi, All-India Survey of Photovoltaic Module Degradation, 2013. Available at: http://www.ncpre.iitb. ac.in/research/pdf/All_India_Survey_of_ Photovoltaic_Module_Degradation_2013.pdf. [7] M. Ko¨ntges, S. Kurtz, C. Packard, U. Jahn, K.A. Berger, K. Kato, T. Friesen, H. Liu, M.V. Iseghem, Review of Failures of

[8]

[9]

[10]

[11]

[12]

[13]

[14]

Photovoltaic Modules, 2014. Available at: http:// iea-pvps.org/fileadmin/dam/intranet/ExCo/ IEA-PVPS_T13-01_2014_Review_of_Failures_of_Photovoltaic_Modules_Final.pdf. M. Ko¨ntges, G. Oreski, U. Jahn, M. Herz, P. Hacke, K.-A. Weiss, Assessment of Photovoltaic Module Failures in the Field, 2017. Available at: http://www.iea-pvps.org/index. php?id¼92&eID¼dam_frontend_push&docID ¼4059. Brij, Field Issues in Crystalline Solar PV Modules. Available at: http://www.brijfootcare.in/ module-field-issues/. IEA-PVPS, PV-failure Survey, 2016. Available at: http://www.iea-pvps.org/index.php?id¼344 &eID¼dam_frontend_push&docID¼3948. Honsberg, C. and Bowden, S. Anti-Reflection Coatings. Available at: http://www.pvedu cation.org/pvcdrom/anti-reflection-coatings. J.H. Wohlgemuth, P. Hacke, N. Bosco, D.C. Miller, M.D. Kempe, S.R. Kurtz, Assessing the Causes of Encapsulant Delamination in PV Modules, IEEE, 2016. L. Dunn, M. Gostein, B. Stueve, Literature Review of the Effects of UV. Exposure on PV Modules, 2013. Available at: https://energy.gov/ sites/prod/files/2014/01/f7/pvmrw13_ps5_aton ometrics_dunn.pdf. TRAXLE, Solar Photovoltaic panels for very hot areas. Available at: http://www.solar-trackers. com/specific-pv-applications/high-temperaturepv-panels.

8: C RYSTALLINE S ILICON PV M ODULE F IELD FAILURES

[15] J. Pern, Module Encapsulation Materials, Processing and Testing, 2008. Available at: https:// www.nrel.gov/docs/fy09osti/44666.pdf. [16] S. Yang, K. Whitfield, Thermal Endurance Study of Polymers Used in Low Concentration Photovoltaic Modules, vol. 2, 2012. [17] D.C. Jordan, S.R. Kurtz, Photovoltaic Degradation RatesdAn Analytical Review, vol. 21, 2013, pp. 12e29. [18] G. Oreski, K. Mo¨ller, Qualification of polymeric components for use in PV modules, in: Proceedings SPIE - The International Society for Optical Engineering, 8112(10), 2011, 81120Be81120B14. [19] F. Pern, Factors that affect the EVA encapsulant discoloration rate upon accelerated exposure, Sol. Energy Mater. Sol. Cells 41e42 (1996) 587e615. [20] A. Fell, K.R. McIntosh, Determining the generation profile for silicon solar cells from lumped optical parameters, in: Proceedings IEEE 42nd Photovoltaic Specialist Conference (PVSC), 2015, pp. 1e5. [21] M. Ko¨ntges, S. Altmann, T. Heimberg, U. Jahn, K.A. Berger, Mean degradation rates in PV systems for various kinds of PV module failures, in: Proceedings 32nd EUPVSEC, 2016, pp. 1435e1443. [22] K. Morita, T. Inoue, H. Kato, I. Tsuda, Y. Hishikawa, Degradation factor analysis of crystalline-Si PV modules through long-term field exposure test, in: Photovoltaic Energy Conversion, 2003. Proceedings of 3rd World Conference on, vol. 2, 2003, pp. 1948e1951. [23] D. Chianese, A. Realini, N. Cereghetti, S. Rezzonio, E. Bura`, A. Bernasconi, Analysis of weathered c-Si PV modules, in: Proceedings 3rd World Conference On Photovoltaic Energy Conversion, 2003, pp. 2922e2926. [24] T. Shioda, 2013 NREL Workshop, 2013. Available at: https://www1.eere.energy.gov/solar/ pdfs/pvmrw13_openingsession_shioda_mitsui. pdf. [25] J. Wohlgemuth, T. Silverman, D.C. Miller, P. McNutt, K. Michael, M. Deceglie, Evaluation of PV Module Field Performance, IEEE, 2015. [26] PVEducation, Hot Spot Heating. Available at: http://www.pveducation.org/pvcdrom/modules/ hot-spot-heating.

215

[27] J. Gallon, G.S. Horner, J.E. Hudson, L.A. Vasilyev, K. Lu, PV Module Hotspot Detection, 2015. Available at: https://www.nrel. gov/pv/assets/pdfs/2015_pvmrw_19_gallon.pdf. [28] DUPONT, Mitigating Strategies for Hot Spots in Crystalline Silicon Solar Panels. Available at: http://www.dupont.com/content/dam/dupont/ products-and-services/solar-photovoltaic-materials/solar-photovoltaic-materials-landing/documents/hot-spot-mitigation.pdf. [29] J.C. Larue, E. d Trieu, Effect of Partial Shadowing on Solar Panels Hot Spot or Breakdown?: Photovoltaic Solar Energy Conference, 1981. [30] P. Hacke, K. Terwilliger, R. Smith, S. Glick, J. Pankow, M. Kempe, S.K.I. Bennett, M. Kloos, System Voltage Potential-induced Degradation Mechanisms in PV Modules and Methods for Test, IEEE, 2011. [31] Wikipedia, Potential Induced Degradation. Available at: https://en.wikipedia.org/wiki/ Potential_induced_degradation. [32] W. Luo, Y.S. Khoo, P. Hacke, V. Naumann, D. Lausch, S.P. Harvey, J.P. Singh, J. Chai, Y. Wang, A.G. Aberle, S. Ramakrishna, Potential-induced degradation in photovoltaic modules: a critical review, Energy and Environmental Science (2017). [33] Fraunhofer_ISE PID e Analysis and Mitigation. Available at: https://www.ise.fraunhofer.de/content/ dam/ise/de/documents/infomaterial/brochures/ photovoltaik/16_en_ISE_Flyer_PID.pdf. [34] B. Weinreich, B. Schauer, S. Seidl, E. Schubert, R. Haselhuhn, Feldstudie zur Modul- und Generatorqualita¨t au f Basis thermogra fischer Messungen u¨ber 100 MW, 2013. Available at: https:// www.solarschmiede.de/sites/default/files/ Solarschmiede-Thermo-Feldstudie-2013.pdf. [35] Firstgreen, Some Common Causes of Solar PV Module Failure, 2014. Available at: http://www. firstgreen.co/2014/07/common-causes-of-solarpv-module-failure/. [36] W. Gambogi, Performance and Durability of Photovoltaic Backsheets and Comparison to Outdoor Performance, 2013. Available at: https:// www.nist.gov/sites/default/files/documents/ el/building_materials/Gambogi.pdf. [37] Wikipedia_Noryl, Noryl. Available at: https:// en.wikipedia.org/wiki/Noryl.

216

D URABILITY AND R ELIABILITY

OF

P OLYMERS

[38] Enclosurecompany, IP Rated Enclosures Explained. Available at: http://www.enclosure company.com/ip-ratings-explained.php. [39] gutachten Bilder aus meiner Arbeit als Sachversta¨ndiger. Available at: http://www.gutachten. streib.de/bilder/index.html. [40] SunPower, SunPower Limited Product and Power Warranty for PV Modules, 2012. Available at: https://www.diablosolar.com/docs/partners/ sunpower-warranty-modules-v2014-8.pdf. [41] Canadian Solar, Limited Warranty Statement Photovoltaic Module Products, 2017. Available at: https://www.canadiansolar.com/downloads/ warranties/en/Canadian_Solar-PV_Module_ Warranty-en.pdf. [42] ReneSola, Limited Warranty Terms For PV Modules, 2014. Available at: https://www.irish ellas.com/files/Warranty-Terms.pdf. [43] Kyocera, Limited Warranty for Kyocera Photovoltaic Module(s), 2015. Available at: http:// www.kyocerasolar.com/pdf/warranty.pdf. [44] Trinasolar, Limited Warranty for Trina Solar Brand Crystalline Solar Photovoltaic Modules, 2015. Available at: http://static.trinasolar.com/ sites/default/files/EN_Warranty_2015_04.pdf. [45] SunEdison, Photovoltaic Module Limited Warranty, 2016. Available at: https://www.energy matters.com.au/wp-content/uploads/2016/12/ warranty-sunedison.pdf. [46] A. Pozza, T. Sample, Crystalline Silicon PV Module Degradation after 20 Years of Field Exposure Studied by Electrical Tests, Electroluminescence, and LBIC, 2016. [47] C. Raupp, C. Libby, S. Tatapudi, D. Srinivasan, J. Kuitche, B. Bicer, G. Tamizhmani, Degradation Rate Evaluation of Multiple PV Technologies from 59,000 Modules Representing 252,000 Modules in Four Climatic Regions of the United States, IEEE, 2016.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

[48] J. Singh, J. Belmont, G. Tamizhmani, Degradation Analysis of 1900 PV Modules in a Hot-dry Climate: Results after 12 to 18 Years of Field Exposure, IEEE, 2013. [49] K. Yedidi, S. Tatapudi, J. Mallineni, B. Knisely, J. Kutiche, G. Tamizhmani, Failure and Degradation Modes and Rates of PV Modules in a Hotdry Climate: Results after 16 Years of Field Exposure, IEEE, 2014. [50] S.B. Darling, F. You, T. Veselka, A. Velosa, Assumptions and the levelized cost of energy for photovoltaics, Energy and Environmental Science (2011). [51] J. Wohlgemuth, Use of Field Observations to Assess PV Module Reliability, 2015. Available at: https://www.nist.gov/sites/default/files/ documents/el/building_materials/Use-of-FieldObservations-to-Assess-PV-Module-ReliabilityJohn-Wohlgemuth.pdf. [52] M. Chicca, J. Wohlgemuth, G. TamizhMani, Comparative Study of 28 and 18 Years Field Aged Siemens-Arco M55 Modules in Temperate and Hot-dry Climates, IEEE, 2016. [53] G.M. Kimball, S. Yang, A. Saproo, Global Acceleration Factors for Damp Heat Tests of PV Modules, 2016. Available at: https://gregory kimball.files.wordpress.com/2012/03/kimball 2016-global-acc-_-presentation.pdf. [54] L.A. Escobar, W.Q. Meeker, A Review of Accelerated Test Models. Statistical Science, 2006.

Further Reading [1] EnergyEfficiency-solar, Photovoltaic Systems Using Micro Inverter?, 2010. Available at: http:// engineering.electrical-equipment.org/energy-effi ciency-solar/photovoltaic-micro-inverter.html.

9 Degradation Processes and Mechanisms of PV Wires and Connectors Sumanth Varma Lokanath 1, Bryan Skarbek 1 and Eric John Schindelholz 2 1 First Solar Inc, Mesa Arizona, United States; 2 Sandia National Laboratories, Albuquerque, New Mexico, United States

9.1 Introduction Photovoltaic (PV) power plants and their constituent components, by virtue of their application, are exposed to some of the harshest outdoor terrestrial environments. Most equipment is subject directly to the environment and myriad stresses (micro and macro environment). Other aspects including local site conditions, construction variability and quality, and maintenance practices also influence the likelihood of such hazards. Many discrete components, including PV modules, combiner boxes, protection devices, inverters, and transformers, make up the PV generation system. The connections between these discrete components are accomplished using PV connectors, PV cables (both above and below ground wires), and wire splices, as shown in Fig. 9.1A and B. Fig. 9.1A illustrates the DC side of a power plant and Fig. 9.1B illustrates the AC side of the power plant (commercial or utility scale). There can be a sizeable quantity (thousands to millions) of connectors or feet of cable that can scale according to the PV power plant size and layout. Wire management devices are yet another family of components used to secure these electrical cables and connectors to mechanical and structural members supporting the PV components. ]Failure of these interconnecting components can affect the generation capacity of the power plant by a factor of the load it is carrying. Of more serious concerns are the safety (e.g., fire and shock) hazards such failures can precipitate. Failures in such components can result in open circuits, short circuits, faults, and leakage driven impacts. Faults can further be classified as ground faults, line faults, and arc faults [1]. Table 9.1 illustrates the potential risk/impact of these failures. A sensitivity study of failure rates completed as part of this study on the DC segment of the system is quite revealing with respect to the impact of failure of

components such as PV connectors, harnesses, and fuses. A wire harness [2] is a preassembled connectorized parallel connection of PV cable lengths and splices. The connectors were segregated into their own category for this study. A change factor of 10 in the baseline assumed failure rate (Rank 5) and its resulting impact on system availability and cost of replacement parts were simulated for a 20 MW DC power plant over 30 years of operation, as illustrated in Table 9.2. The simulated data indicate that beyond the PV module, failure rates in cables, cable splices, fuses, and connectors can be significant destroyers of DC health in that order ranked on the Availability or Spare Part Costs. Assumptions for availability include corrective measures implemented on failure occurrence within a 30-day period. From field data, when partitioning causes of field replacements from the harness, w40% of harness replacements were from connector failures, ~25% from wire failures, ~20% from fuses, and balance w15% from installation issues, which also supports the above simulation findings. Wire management can also influence the failure rates and for simplicity is excluded from the analysis. The failure and remediation of these components can however incur a significant labor cost and expose the overall system to additional risks as discussed later in Section 9.7. Another challenging issue with failures of these components is that of fault localization. Depending on the failure location, only a single PV module string may be impacted, and such outages may not be readily detected and remediated with the monitoring infrastructure in place. The expected energy output of the plant (throughput) can be degraded due to component outages and can be expressed as a ratio of actual to expected, and is defined as the throughput capacity ratio. Fig. 9.2 illustrates the throughput capacity ratio impacted by detection and repair delay days.

Durability and Reliability of Polymers and Other Materials in Photovoltaic Modules. https://doi.org/10.1016/B978-0-12-811545-9.00009-4 Copyright © 2019 Sumanth Varma Lokanath & Bryan Skarbek. Published by Elsevier Inc. All rights reserved. Contribution by Eric John Schindelholz is in public domain.

217

218

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 9.1 (A) Illustration of components in the DC side of a PV power plant. (B) Illustration of components in the AC side of a PV system.

Table 9.1 Potential Risk/Impact of Failures Risk Performance

Shock

Fire

System Shut Down

Open circuit

Yes

Possible

No

Possible

Short circuit

Yes

Possible

Possible

Possible

Ground fault

Yes

Possible

Possible

Possible

Line fault

Yes

Possible

Possible

Possible

Arc fault

Yes

Possible

Possible

Possible

Leakage

Possible

Possible

Possible

Possible

Failure

9: D EGRADATION P ROCESSES AND

Failure Rate

Spare Part Costs ($)

Availability (%)

Rank

10

1

1

1

61,893,187

97.4

1

1

1

10

1

6,840,058

99.42

2

1

1

1

10

1,769,962

99.66

3

1

10

1

1

1,162,080

99.84

4

1

1

1

1

1,131,350

99.85

5

1

0.1

1

1

1,115,021

99.86

6

1

1

1

0.1

1,087,949

99.87

7

0.1

1

1

1

677,779

99.92

8

1

1

0.1

1

643,421

99.89

9

Bold highlights the top 3 riskiest components by Rank.

C ONNECTORS

Fuse

AND

Harness (Wire D Splice)

PV W IRES

Connector

OF

PV Module

M ECHANISMS

Table 9.2 DC Field Sensitivity Study Failure Rates

219

220

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 9.2 Throughput capacity ratio impact due to fault detection and repair delay times.

Additionally, there could be an impact on module degradation during the time the string is in open circuit. This is seen in all curves except the 180-day curve where it is not a significant factor due to time. Fault detection and remediation should therefore occur in a timely manner to minimize losses. While there are abundant data that illustrate PV modules and PV inverters to be the major contributors of PV system failures, the mentioned data illustrate the importance in minimizing failures in the often ignored components such as PV connectors, PV wires (both above and below ground), wire splices, fuses, fuse holders, fuse holder enclosures, and wire management devices. With the exception of PV fuses, these components predominantly use polymeric materials. Therefore, it is crucial to understand the degradation processes and mechanisms leading to component failure and their impact on system performance or failure. The following sections treat each of these components independently and provide an insight into commonly found and emerging failure modes and mechanisms, considerations, and impacts on overall system health.

9.2 PV Connectors PV connectors are traditionally single pole locking connectors used between the DC carrying parts of the PV system, as shown in Fig. 9.3. These are high voltage, high current, and nonload breaking devices and often subjected to voltages as high as the system

voltages (600e1500 V). Recently with the introduction of tracking structures, other multipole connectors that are low voltage, high current, and nonload breaking devices are also finding use in PV systems. Regardless, all connectors in PV applications are outdoor, exposed over the lifetime of the power plant. There are other connectors used with inverters and enclosed assemblies that are not discussed here. The focus in this section pertains to single pole outdoor exposed connectors used to connect the DC energy within the PV system. Fig. 9.3 illustrates the typical construction of a common locking connector used in PV. The body and endcaps of locking connectors are typically constructed of a UV-resistant polycarbonate. Rubber seals are also common to minimize moisture and dirt intrusion. Pins of these connectors are typically copper with tin or silver plating. The plating provides corrosion protection and reduces contact resistance relative to bare, oxidized copper. Material choice is an important consideration. Typical thermoplastic materials used in the construction of PV connectors are polyamide (Nylons), Poly Vinyl Chloride (PVC), and ThermoPlastic Urethane (TPU). When exposed to the elements, thermoplastics are robust polymers but they remain vulnerable to environmental stressors. Although they are cheap and easy to manufacture, thermoplastics may decompose (or depolymerize) when exposed to heat, radiation, and/or oxidants. As with most polymers, thermoplastics may also experience side reactions between polymers and fillers, which degrade

9: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

PV W IRES

AND

C ONNECTORS

221

Figure 9.3 Typical female and male sides of a PV connector with exploded view of components (male and female).

the mechanical properties of thermoplastics over time. Additives and unreacted monomers in thermoplastics may also migrate or diffuse out of the bulk and surface to cause the loss of mechanical stability and integrity. The severity of environmental degradation of thermoplastic polymers is shown in the spider chart in Fig. 9.4 on a scale from zero to two: “0” signifies the polymer is not vulnerable, “1” indicates the polymer is vulnerable, and “2” indicates the polymer is extremely vulnerable and prone to degradation [3e6]. A collaborative field study conducted by First Solar, Case Western Reserve University and Underwriter laboratories (UL) subjected combinations of PV connectors and wires to accelerated indoor exposures of

cyclic and multistress factors, as well as real world outdoor exposures. A Fourier transform infrared spectroscopy (FTIR) analysis of the accelerated exposure samples indicated polymer degradation processes occurring due to UV exposure impacts. Additional data from this research are pending. Degradation of the connector housing and improper installation can lead to an increase in connector contact resistance. These events can increase transport of moisture and environmental contaminants to the metallic electrical contacts, in turn causing corrosion. Differential thermal expansion and contraction caused by diurnal temperature changes could additionally introduce fretting of the connector plating, accelerating the process [7]. Full or partial demating of the

Figure 9.4 Degradation aspects and severity of thermoplastics.

222

D URABILITY AND R ELIABILITY

OF

P OLYMERS

contacts may result from mechanical damage as discussed in the following section. Beyond the obvious, though possibly acceptable, costs of ohmic power loss, connector contact degradation can also lead to arc fault events. There have been instances where arc faults related to PV connectors have been documented with their prevention being identified as a critical knowledge gap [8,9]. In one study, connector failure was attributed to causing 29% of fires in surveyed PV system fires, with module (34%) and other BOS components (37%) making up the remainder of fires [10]. Other than a possible root cause, the risk of arcfault from connector corrosion remains unclear. A recent study found commercial connectors to be relatively resilient to several types of laboratory corrosion tests, including damp heat (85  C/85% RH) with sea salt contaminants. The increase in contact resistance observed during testing was attributed to corrosion but was relatively insignificant with respect to arc fault risk, as shown in Fig. 9.5. The authors state that further work is needed to capture and accelerate contact degradation seen in the field and quantify its impact on arc faults [8].

9.2.1 New Failure ModesdWithdrawal Force The cable to connector interface is subjected to multiple mechanical stresses during the lifetime of 4.5 Type 2 Type 3 Type 1

Resistance (mΩ)

4.0

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

the product. One major failure mode of this interface is the cable pulling out of the connector body, which could potentially lead to a safety hazard (shock or fire). The force at which this occurs is named the withdrawal force. Steps should be taken to ensure that the withdrawal force is high enough to withstand the stress that the connector could see in the field. Some possible causes that may result in the cable being extracted from the connector body result from higher than normal forces due to the added weight of the bundled harnesses. Examples include: 1. cable ties or other means of supporting large bundles of cable failing with the connector taking this additional load; 2. snow/icedif the cable is locked in place by snow or ice and the tracker starts to track, this could put high stress at the connector to cable interface; and 3. improper wire managementdif the cable is not properly routed, it can get hung up on moving parts. The locked connector release force test ensures that in the event of a wire management failure, the PV connector will release at the male to female interface instead of the cable pulling out of the body of the connector. Ideally, the locked connector release force should be greater than 89 N (20 lbf) but less than the withdrawal force. If samples are successfully tested (tensile testing) to these criteria, this will lower the risk of having exposed conductors in the field.

9.2.2 Considerations for PV Connector Factory versus Field Assembly

3.5 3.0 2.5 2.0

AND

0

2000

4000 6000 8000 Elapsed Time (hours)

10000

Figure 9.5 Resistance measured across three models of mated commercial quick connectors during damp heat (85  C/85% RH) testing. Connector pins were contaminated with sea salt (dotted line), simulated desert dust (dashed line), or left in asreceived condition prior to testing. From Schindelholz (2014).

Field failure data from several installations for w1.9 GW DC for 1 year are plotted in Fig. 9.6. These data are within the initial life period of the system (1e3 years). As seen, connector failures are the second item on the Pareto chart after fuses, with an order of magnitude lower. The failure rate is in the 0.01%e0.1% order of magnitude. These low numbers are attributed to controlled factory assembly of connectors. The assembly of connectors onto cables occurs in a factory setting under controlled conditions with a formal and robust quality assurance program.

OF

PV W IRES

e ul

ba

od

3

2

2

1

1

1

H

M

C M W

6

r H C B fu se ar ne Fu ss se ba H d ol de r D H is C co B N nn ot ec Te tio rm n in at ed in ... Fe ed er

14

d

4

20

hi

d Ba

M ECHANISMS

O th e

33

p

450 400 350 300 250 200 150 100 50 0

AND

Pareto of “Harness and HCB” Failures over 1 Year

417

IL F

Quantity

9: D EGRADATION P ROCESSES

Figure 9.6 Pareto of “harness and HCB” failures over 1 year.

Table 9.3 illustrates the critical considerations for quality control in the connector assembly process, and its impact on failure modes. Such considerations can be significant control factors/variables. There are human factors issues to consider in a field installed environment that introduce a wider variance of noise into control factors/variables as illustrated in Table 9.3. For example, manual installation of connectors is repetitive (hand fatigue) and error prone (wasted and faulty parts). Crimp quality variation is difficult to control and dust and debris can influence connection quality or reliability. Further, detection of bad quality assemblies is not possible in a field setting. Therefore field assembled connectors are inherently expected to demonstrate higher failure rates and can potentially lead to increased open circuit, exposed live parts (shock hazard), excessive leakage current (shock hazard),

AND

C ONNECTORS

223

and high resistance/arcing (fire hazard). The authors of this chapter estimate the field assembled PV connector failure rate in the order of approximately 2%e5% as surveyed by O&M practitioners. Therefore, factory assembly of PV connectors (PV source and branch circuits) is strongly recommended. Alternatively, strict control of assembly in the field is necessary, such as proper use of tools, calibration of equipment, and training certification. If field replacement of factory-assembled connectors is necessary, the authors suggest considering field assembly of preassembled connectorized ends with butt splices (see Fig. 9.7). Crimping a butt splice is simpler (see Fig. 9.8) and reliable than crimping and assembling PV connectors with multiple parts in the field. Further, an approved insulation material can be easily applied to insulate the splice. The long-term durability of heat shrink (see Fig. 9.9) is an important consideration. Alternatively, an additional cover may be considered in addition or in lieu of heat shrink tubing. Finally, proper wire management and strain relief for spliced joints is another aspect to address in order to ensure reliability.

9.3 PV Wire (Above Ground) Chlorinated Polyethylene (CPE) and HighDensity Polyethylene (HDPE) are thermoplastic materials whereas Ethylene Propylene Diene Terpolymer (EPDM) and Ethylene Propylene Rubber (EPR) are elastomers; Cross-Linked Polyethylene

Table 9.3 Summary of Critical Considerations in Connector Assembly Process Requiring Quality Control Measures Steps in the Assembly Process that Need Special Quality Control

Result Too Short/Low

Too Long/High

Strip length

Pins cannot be fully inserted into the body

Exposed conductor/Excessive leakage current

Crimp height

Deformation of metal components

Low withdrawal force

Fire risk/high resistance; low withdrawal force

NA

Fire risk/high resistance

NA

Excessive leakage current

Excessive leakage current

Dirt and moisture can lead to corrosion on internal parts

NA

Pin fully inserted into the body Mating area across pins Torque on cap nut Cleanliness of components

224

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Figure 9.7 Illustration of a preassembled connectorized end with butt splice for field use. A single crimp in field using standard tools.

Figure 9.8 Illustration of a butt splice connector with one end crimped in factory.

(XLPE) is a thermoset. Polymers used in the construction of PV wires typically use EPDM, XLPE, or EPR/EPDM composite materials for the insulation and CPE or composite HDPE/XLPE for the sheathing or jacket. Thermoplastics are already described in Section 9.2. Elastomers may depolymerize when exposed to heat, radiation, and/or oxidants. Elastomers and especially natural rubbers are more prone to corrosion than any other class of polymers and they are particularly vulnerable to ozone, moisture, bacteria, and fungi. Rubbers are also extremely flammable. As

Figure 9.9 Heat shrink applied to cover the butt splice connector in field.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

with most polymers, elastomers may age mechanically over long periods and degrade by experiencing phase segregation or side reactions that are caused by polymer chain scission or crosslinking [11,12]. Thermosets are relatively more resilient to the elements when compared to thermoplastics and elastomers. Thanks to their three-dimensional network of covalent bonds, thermosets exhibit great mechanical and thermal stability when exposed to the elements. Despite their thermo-mechanical stability, thermosets are extremely prone to oxidation, photolysis (radiation-induced depolymerization), and side reactions with fillers and additives. Thermosets are prone to physical and mechanical aging over long periods and may experience phase segregation to some degree as mass diffusion paths may be hindered by crosslinks [13,14]. Polymer composites are mixtures of different polymers in a single-phase material. Composites typically experience hybridized properties that originate from their constituent polymers. Blending polymers may not only be used to enhance their mechanical and chemical stability but they may also be used to reduce the cost of polymer components, as is the case with EPR/EPDM. They both have a similar structure but different properties. EPDM is more durable against heat, oxidation, and mechanical aging than EPR, thanks to its saturated bonds, but EPDM is also much more expensive than EPR. Hence by mixing EPDM and EPR, a trade-off is made between enhanced properties and reduced costs. With the HDPE/XLPE combination, a thermoset and thermoplastic give the best of both worlds, and mixing low-cost HDPE with high-cost XLPE balances the price of components. Although composites and nanocomposites are typically better performers than their constituent polymers, they are still vulnerable to environmental degradation and may depolymerize when exposed to heat, radiation, or an oxidizing environment. Composites are especially vulnerable to moisture, bacteria, mechanical aging, phase segregation, and side reactions between the different constituent polymers, additives, fillers, and substrates. The severity of environmental degradation of elastomeric polymers is shown in Fig. 9.10 on a scale from zero to two, from least to most severe [15e20]. There is no right or universal solution in selecting polymers for PV components as the decision depends mainly on the physical and chemical composition of the environment at hand. Rubber composites such as

9: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

PV W IRES

AND

C ONNECTORS

225

Figure 9.10 Degradation aspects and severity of polymers used in PV cables.

EPR/EPDM may be a cost-effective solution in dryer environments with high UV radiation such as the Arizona desert. Other composites such as HDPE/ XLPE may be a great choice for components in humid environments, with slightly less UV radiation such as the Dubai desert. Polymer material selection must be designed around the presence of UV photons (primarily), oxidants, heat, and long-term mechanical aging, which are typically the main modes of polymer degradation.

9.3.1 New Failure ModesdUV Robustness UV robustness in sunlight-exposed cables is commonly achieved utilizing additives of carbon black and this provides the black color for the cable. However, there is also a trend where the UV robustness is alternately provided by addition of light stabilizers such as Tinuvin and a coloring agent is added to give the black color or any alternative color desired by the customer. There are limited field reported data that the latter form of cables sees premature cracking failures in the field. The dosing quantity of light stabilizers may be a contributor. A lower flexibility conductor may further exacerbate the issue.

9.3.2 New Failure ModesdSlip Force Cables from end of PV module strings or a parallel combination of strings leading to combiner boxes are called as PV subarray cables. PV subarray cables are often terminated in fuse holders within combiner boxes. Such PV subarray cables may also transition from above ground to below ground using connectorized distinct sections with the other end terminated directly at a fuse terminal in a combiner box. Thermal cycling stress from diurnal temperature changes may result in the shrink-back of the insulation, which has been found on some PV cables. This shrink-back is the result of the cable insulation “walking” over the cable due to these thermal cycles. Shrink-back can potentially be a safety hazard or can lead to a shock or thermal event if the cable is terminated to a fuse holder or terminal block and the insulation creeps back far enough to expose a large area of the conductor as shown in Fig. 9.11. If the exposed conductor makes contact with grounded metal or conductors of opposite polarity, this could result in arcing. Preliminary data indicate that some insulation materials, such as XLPE, are more susceptible to shrink-back than other materials when

226

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Figure 9.11 Cable shrink-back exposing bare conductor on fuse terminals.

exposed to thermal cycling. Further research is necessary to determine if this failure requires to be alleviated by altering material properties, but at this time there has been no major field failures reported due to shrink-back.

9.3.3 New Failure ModesdRoundness of Cables (Filled vs. Unfilled) Multiconductor cables are commonly used for communication signals on tracker systems in PV plants. One of the biggest challenges is preventing moisture and dust ingress at the connector to cable interface on these harnesses due to the roundness of the cable. Connectors typically use a gland with a back-nut that is torqued to the manufacturer’s specifications. The proper torque value is critical to ensure that a uniform clamping force is applied to the gland. If an unfilled cable is used, the shape of the outer diameter is more of an ellipse shape instead of it being round. This allows voids between the gland and the insulation of the cable that can allow moisture ingress. The fix is to use a filled cable, which maintains a symmetric cable shape and prevents voids between the gland and the insulation of the cable.

9.3.4 New Failure ModesdCable Subjected to Flexing Degradation of the cable and conductor insulation can occur due to cyclic bending or flexing in the field especially when installed on PV trackers. PV trackers go through cyclic movement every day and can

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

further be impacted by random cyclic movements from high wind effects. Cables and assemblies should be carefully chosen to be robust to such stresses as polymeric materials are impacted and may possibly result in leakage, shock, and fire hazards. Further laying cable on sharp edges, pinch points, and introducing twists in cable leads to the fatigue of the conductor and shield in the flex area. Fatigue of the conductor and shield at the point of termination (stress at the termination point) may lead to individual stands of the cable failing, or the grounded shield becoming isolated from the ground. The appropriate flexibility classification should be considered when installing cable sections on tracker systems or other systems with moving part as described in the next section. The choice of material type in the application also matters. EPR insulation may be suited better than XLPE for hot environments, and where flexibility is critical. Further, additional steps can be taken to address this issue, such as conducting mockups and detailed reviews prior to the field rollout, through proper training during first installations, and verified by evaluation through compressed time Accelerated Life Testing (ALT). Finally, all the learning is to be documented in detailed wire management drawings and installation instructions.

9.3.5 Approach for Determining Bend RadiusdVarious Competing Sources A key failure mode for wires is the insulation degradation from stress caused due to wire management or the lack of it. Wire management requires that best practices and measures for installation be mandated, such as avoiding: needless connections, sharp twists in cable, transitioning/laying the cable over rough/sharp edges, tight spaces, thermal expansion, and choosing cables with the appropriate flexibility class. The flexibility of the cable is a critical consideration depending on the application scenario as illustrated in IEC 62440 [21]. There are differences in the minimum bend radius criteria which lead to the dilemma of which one to choose. The cable manufacturers may have their own compliance requirements for bend radius to assure warranty and service life. Table 9.4 provides a rational waterfall-based approach to use for demonstrating compliance. It is important to note that PV connector manufacturers may also impose bend radius criteria on

9: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

PV W IRES

AND

C ONNECTORS

227

Table 9.4 Summary of Wire Management Guidance and Condition Based Priority Priority

Bend Radius Governance

Guidance

Reference

1

Use manufacturer guidance

Use this, especially if contingent to secure warranty/service life per manufacturer

Use dynamic for tracker use

1

International Electrotechnical Commission

IEC 62548:2016 Section 7.3.7.3 [21]

Manufacturer guidance

2

National Electric Code

NEC 338.24 for USE-2 cable type

5  OD

2

International Electrotechnical Commission

IEC 62440 Section 5.6 Table 9.3 a [21] 4  OD (free movementdmodule to module) 6  OD (mechanical loaddend of module string)

4  OD 6  OD

a

Application should also consider a fixed tilt use versus a tracker installation. While IEC 62440 [21] applies for lower voltage cables (450e750V), a similar approach of implementation is reasonable for higher voltage PV cables.

cable transitions terminating inside connectors. It is also equally important to consider maintaining the minimum bend radius of Outer Diameter (OD) of cables in packaging, shipment, and handling of cables or premanufactured harness assemblies, from the supplier all the way to the installation site. Flexibility class is interpreted differently in various regions. The IEC 60228 [21] standard defines flexibility class by the cross section of the conductors, whereas within the United States flexibility class is governed by the strand count. The vast majority of systems in the United States use a 19-strand cable and there have been no reported gross failures with this cable. US installations also use a single insulated cable whereas a double insulated cable is required within the IEC countries. There has been recent trend to require flexibility class 5 for PV cables. While the authors believe there is justified use for special requirements, over generalization of such requirements will needlessly increase the costs of PV cables. Fixed installations and even flexible areas where the cables are managed and supported do not require Class 5 cables. Class 5 cables are recommended in areas subjected to unrestrained loading/twisting or movement such as tracker to fixed transitions as described in IEC 62738 Section 7.3.4.4 [21].

9.4 PV Wire (Below Ground) Deterioration or damage to wire insulation of buried cabling could lead to leakage current through the soil that is capable of corroding metallic PV module mounting systems and nearby third-party infrastructure (e.g., pipelines). This risk, especially for large PV plants with kilometers of buried cable, has recently been brought to attention [9,22]. In PV installations, the leakage current loop will flow through the ground via parallel paths of least resistance. Metallic structures buried or in contact with the ground can serve as preferential least resistance paths, given their conductive nature, and conduct this current. At each point where current leaves the metallic structure to earth on this path enhanced corrosion can occur (Fig. 9.12). Although PV systems incorporate ground fault detection, Charalambous and coworkers argue this leakage current could go undetected [22]. For example, in a 300 kW grounded PV system, the fault current detection threshold to prevent catastrophic failures could be as high as 5 A. Although documented cases of PV stray current corrosion are lacking in the literature, calculations by Charalambous and coworkers suggest that it should be taken into consideration during design and operation over 25 year

228

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 9.12 Hypothetical schematic of leakage current from cabling causing stray current corrosion in metallic pipeline [23].

lifetimes of commercial PV systems [22]. Furthermore, stray current corrosion is a common degradation mode of buried infrastructures, such as pipelines, especially in the area of linear electrified rail lines [23].

9.4.1 New Failure ModesdTermite Resistant Cables A termite resistant cable is required in some jurisdictions for any direct burial cable. The most common methods used to achieve “Termite Resistant” classification are to either chemically treat the insulation of the cable or to add a Teflon or Nylon material between the conductor and the insulation of the cable. There are concerns to address with both methods. The chemically treated cable requires wearing gloves when handling the product and can put off a strong odor when storing the cable indoors or in shipping containers. Airing of this material is recommended to give the odor time to dissipate and not affect the installation crews. On the other hand, the Teflon or Nylon material can reduce the insulation slip force and allow the insulation of the cable to slide on the conductor. This could potentially be a safety hazard if the cable is terminated to a fuse holder or terminal block and the insulation creeps back far enough to expose the conductor. If the exposed conductor is in close proximity to the grounded metal or conductors of opposite polarity, this could result in arcing and possibly a thermal event. An approach to arrest the cable from the slip force is to utilize the preassembled connectorized

ends with butt splices as described earlier in Section 9.2.2.

9.5 Wire Splices and In-Line Fuse Holders A wire harness is a factory assembly component that aggregates the output of multiple PV string conductors along a single main conductor. Wire splices and in-line fuses are components of wire harnesses; they are typically constructed of an overmold and an under-mold material. The over-mold material must be robust enough to withstand the environmental stress that it will be exposed to for 25 years. Cold impact and UV exposure tests are commonly used to determine if the polymer can withstand the time. As for the under-mold, it is critical that the material makes a good bond to the insulation of the cable to ensure there is no moisture or dust ingress into the component. A presence of moisture within the joint or in-line fuse can lead to corrosion or excessive leakage current. Currently, the UL9703 standard covers such assemblies. The authors found that the IEC 61215 [21] damp heat exposure test was found to identify field correlated failure modes that the UL standard does not address. Below are further recommendations for inclusion in the technical specification requirements for wire splices and in-line fuse holders:

9: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

PV W IRES

1. A force of greater than 200 N is needed to disconnect cables from wire splices. 2. It should be ensured that dissimilar galvanic metals are not used in wire splices that would form a galvanic couple. 3. The crimp should be rated for both aluminum and copper conductors when splices interface the two. 4. The choice of splice design and construction must consider typical stresses in packaging, shipment, installation, and operation, such as torsional bending, strain from load, and current ampacity. Solder or weld by itself must not form the primary mechanical connection; an alternate mechanical securement should be provided in addition to solder. 5. Heat shrink tubing alone must not be used as primary insulation due to the multicable construction unless appropriately chosen for the design and end use environment. 6. The over-molds should be capable of accepting the cable dimensions and durometers used for the cables to which they are fitted. 7. Cavities should be avoided in splice overmolds. All cavities (e.g., pin location points and windows) must be potted with insulation rated for the environment and additionally form a rigid seal preventing pollution ingress paths. 8. The minimum over-mold coverage length for each wire in splice must be dimensioned to protect insulation damage against the expected bend radius range (temporary and sustained) in normal application. 9. The splice insulation must have a voltage rating equal to, or greater than, the PV system voltage and have a dielectric withstand voltage rating equal to, or greater than, twice the PV system voltage plus 1000 V. 10. The splice insulating material must have a temperature rating of 40 to 90 C. 11. The wire splices, if exposed to the environment, must be rated for outdoor use, be UVresistant, and have a minimum rating of IP65 (IP68 is recommended). 12. Wire splices and other components of a wire harness must prevent moisture ingress.

AND

C ONNECTORS

229

In addition to the product specifications for the construction of components within wire harnesses, there should also be installation instructions that provide guidance on 1. the maximum/minimum bend radius of cables going into wire splices during packaging/ handling, 2. recommended best practices for unpacking and handling during installation, 3. the maximum/minimum bend radius of cables going into wire splices during site installation, and 4. wire management recommendations to prevent damage to wire harness.

9.6 Combiner Box A combiner box is generally a metal enclosure that houses components such as terminals, fuse holders, disconnect switches, surge suppression, metrology, etc. Polymeric materials are often used in these components, and are subject to degradation mechanisms resulting in failures. Doors of combiner box enclosures use rubber gaskets. The gasket materials are prone to mechanical and thermal degradation. Aging allows water ingress into the junction box and can significantly affect the ingress properties of the box. Multiple replacements may be necessary over the lifetime of the plant. Material selection and service lifetime prediction are important considerations. Combiner boxes utilize disconnect switches which expose the polymeric handle to the elements outside of the box. Such handles are prone to UV, humidity, and thermal degradation. Again, material selection and service lifetime prediction are important considerations. Combiner boxes have multiple termination points that are prone to thermal events if the hardware is not properly torqued or the equipment used in the design is not rated for its intended use. Equipment such as fuse holders, terminal blocks, and disconnect switches must be rated for use with other components in the design, and tested as such. Using products outside of their intended use may result in field failures even if they passed testing during the initial qualification. An example of this is using a fuse holder rated for use with PV cables,

230

D URABILITY AND R ELIABILITY

OF

P OLYMERS

but in the application, a buss bar is used to collect the power. The profiles of the terminals on the fuse holder are designed to increase the contact area of the round cable, but the comb profile has square contacts that engage in the fuse holder terminal. This decreases the surface area of the contacts. The stress from thermal cycling in the field may allow the torque on the hardware to back off over time and increase the contact resistance, eventually leading to a thermal event as shown in Fig. 9.13. Loosening of fasteners over time can lead to catastrophic failures, and the choice of insulation materials can impact a terminal’s thermal robustness to torque loosening issues. A thorough design review and lab testing can catch these failures during the qualification stage. It is recommended to subject samples of combiner boxes to an Accelerated Life Testing (ALT) of thermal cycling test from 40 to 55 C for 1000 cycles. In addition to the thermal cycling preconditioning, power cycling the unit during the high and low temperature dwells will give a true representation of field conditions as illustrated from an experiment conducted by authors. Post thermographic imaging and resistance measurement tests will identify any hot spots that may result in an increase in contact resistance. Fig. 9.14 shows the temperature rise of terminals within a combiner box that were not torqued to the correct value with temperature and power cycling.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

9.7 Wire Management Devices Millions of feet of wires can span the length and width of the structures supporting the PV modules in a power plant. Wire management of such cables becomes important and in most cases it is also governed by requirements in codes and standards. Polymeric materials of the Nylon type are commonly used for wire management products such as wire ties, clips, and tethers. Nylon 6 and specialized blends of Nylon 6 called Nylon 6.6 are typically used for cost effectiveness. Such materials are susceptible to stress corrosion cracking (SCC) when applied on galvanized steel structures. Run off Zn from the galvanized steel [24] is one of the leading drivers of salt embrittlement failure mode in Nylon 6. Results indicate that a critical Zn ion concentration window of 0e2000 ppm of Zn is required to facilitate this failure. A screening test utilizing ZnCl was developed [24]. Additionally, due to regulatory policies, site preparation and site management chemicals such as surfactants, dust inhibitors, and weed inhibitors are often utilized. The polymeric material to be used on a PV site should be verified to be robust against the exposure to such site chemicals at the concentrations typically used on site. Lastly, acid rain and naturally occurring salts are additional considerations. First Solar utilizes a custom test profile called as Universal Chemical Exposure Test (UCET) developed by its

Figure 9.13 Fuse holder overheating on a bus bar interface.

9: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

PV W IRES

AND

C ONNECTORS

231

Thermal Measurements vs. Power / Temperature Cycles 400

50

350

200

-100

150 -150

100 50 0

Cycle Temperature C

-50

250

-200

1 191 381 571 761 951 1141 1331 1521 1711 1901 2091 2281 2471 2661 2851 3041 3231 3421 3611 3801 3991 4181 4371 4561 4751 4941 5131 5321 5511 5701 5891 6081 6271 6461 6651 6841 7031 7221 7411 7601 7791 7981 8171 8361 8551 8741 8931

Response Temperature C (Dissipiaon)

0 300

-250

-50 -100

-300

Number of Cycles Measured Temperature (C)

Cycle Temperature (C)

Figure 9.14 Temperature rise of a terminal increases after multiple temperature/power cycles.

Reliability Engineering team to validate polymers against such field occurring chemicals. Fig. 9.15 illustrates that the Nylon 12 material has a lower variability response to the max failure loads in tensile testing post environmental exposures as identified by sequences. Also all failures for Nylon 12 material were attributed to Tensile Load (TL) failures in contract to Nylon 6.6 where failures from Chemical Exposures (CE) were also observed. Subsequently, based on testing as illustrated in Fig. 9.15, Nylon 12 has a better performance to chemical robustness, temperature variations, and high humidity conditions [24]. The initial upfront cost is a major decision factor when procuring the millions of cable ties that are

installed in PV plants. However, the authors contest that with labor accounting for more than 97% of the cost to install or replace cable ties in the field, it is more cost effective to pay the higher upfront cost for a product with a longer lifespan. When the authors compared the total cost of ownership over the 25year plant life, paying the higher upfront cost for a more robust cable tie pays off. The cost of Nylon 12 cable ties is three times higher than that of Nylon 6.6 ties but Nylon 12 are expected to last twice the life expectancy than Nylon 6 versions. Therefore, a single replacement cycle can satisfy the design lifetime versus three necessary replacements for Nylon 6 variants. First Solar has switched over to using the superior Nylon 12 as the specified material on its sites because of these findings. The authors strongly urge the PV industry to use Nylon 12 over Nylon 6.6 for PV wire management applications for lifetime cost, performance, and sustainability. PVDF wire ties are currently under development, which may only require a single installation to meet the 25-year design lifetimes of PV power plants.

9.8 Conclusion

Figure 9.15 Variability plot of wire tie tensile load performance post exposure to test sequence.

In summary, this chapter describes the typical materials used in components, the observed degradation processes and mechanisms leading to component failure, and its impact on system performance or failures. It further provides some practical considerations, approaches, and methods in

232

D URABILITY AND R ELIABILITY

OF

P OLYMERS

addressing the problems with practical solutions in the design to assure the performance of the PV plant over the intended design lifetime.

Acknowledgment The authors (Sumanth Lokanath, Bryan Skarbek, and Eric Schindelholz) would like to acknowledge the valuable contributions of Mr. Rajan Bedi, Mr. Sundar Subramanian, Mr. John Pickens, and Mr. Paul Williams from First Solar for their contributions with root cause of field findings; Dr. Mounir El Asmar and Mr. Georges Nassif from Arizona State University for their contributions in the review of various polymers and corresponding failure mechanisms and developing the spider charts; Dr. Roger French and Mr. Timothy Peshak of Case Western Research University and Mr. Liang Ji from Underwriter Laboratories for their ongoing work on PV connectors and wires. This work was in part supported by the Laboratory Directed Research and Development program at Sandia National Laboratories. Sandia National Laboratories is a multi-mission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DENA0003525.

References [1] A. Chokor, M.E. Asmar, S.V. Lokanath, A Review of Photovoltaic DC Systems Prognostics and Health Management: Challenges and Opportunities - Annual Conference of the Prognostics and Health Management Society, 2016. [2] Harness Assembly as Defined in Standards UL9703, IEC 62780 [12]. [3] Kambour, A review of crazing and fracture in thermoplastics, Journal of Polymer Science Macromolecular Reviews 7 (1973) 154. [4] Allen, Chirinis-Padron, Henman, The photostabilisation of polypropylene: a review, Polymer Degradation and Stability 13 (1) (1985) 31e76. [5] D. Gavrila, E. Gosse, Post-irradiation degradation of polypropylene, Journal of Radioanalytical and Nuclear Chemistry 185 (2) (1994) 311e317.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

[6] S. Satoto, T. Yusiasih, Watanabe, Hatakeyama, Weathering of high-density polyethylene in different latitudes, Polymer Degradation and Stability 56 (3) (1997) 275e279. [7] Y.W. Park, T. Sankara Narayanan, K.Y. Lee, Fretting corrosion of tin-plated contacts, Tribology International 41 (7) (2008) 612e628. [8] E. Schindelholz, et al., Characterization of fire hazards of aged photovoltaic balance-of-systems connectors, in: Photovoltaic Specialist Conference (PVSC), 2015 IEEE 42nd, IEEE, 2015. [9] A. Demetriou, et al., Stray current DC corrosion blind spots inherent to large PV systems fault detection mechanisms: elaboration of a novel concept, IEEE Transactions on Power Delivery (2016), 2016. [10] M. Kontges, S. Kurtz, C. Packard, U. Jahn, K.A. Berger, K. Kato, T. Friesen, H. Liu, M. Van Iseghem, Review of Failures of Photovoltaic Modules, International Energy Agency, 2014. [11] A.K. Bhowmick, H.L. Stephens, Handbook of Elastomers, second ed., M. Dekker, New York, 2001. Print. “Plastics Engineering” (Marcel Dekker, Inc.) ; 61. [12] S. Deuri, Bhowmick, Ageing of rocket insulator compound based on EPDM, Polymer Degradation and Stability 16 (3) (1986) 221e239. [13] J.-P. Pascault, Thermosetting Polymers, Marcel Dekker, New York, 2002. Print. Plastics Engineering (Marcel Dekker, Inc.) ; 64. [14] S. Mitra, Ghanbari-Siahkali, Kingshott, Rehmeier, Abildgaard, Almdal, Chemical degradation of crosslinked ethylene-propylene-diene rubber in an acidic environment. Part I. Effect on accelerated sulphur crosslinks, Polymer Degradation and Stability 91 (1) (2006) 69e80. [15] N. Petchwattana, S. Covavisaruch, S. Chanakul, Mechanical properties, thermal degradation and natural weathering of high density polyethylene/ rice hull composites compatibilized with maleic anhydride grafted polyethylene, Journal of Polymer Research 19 (7) (2012) 1e9. [16] R. Rothon, R.N. Rothon, Rapra Technology Limited Content Provider, Particulate-filled polymer composites, second ed., iSmithers Rapra, Shrewsbury, 2003. [17] Osawa, Sunakami, Fukuda, Photodegradation of blends of poly(vinyl Chloride) and polyurethane, Polymer Degradation and Stability 43 (1) (1994) 61e66.

9: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

PV W IRES

[18] O. Haillant, Spectroscopic characterization of the stabilising activity of migrating HALS in a pigmented PP/EPR blend, Polymer Degradation and Stability 93 (10) (2008) 1793e1798. [19] Choudhury, Bhowmick, Ageing of natural rubber-polyethylene thermoplastic elastomeric composites, Polymer Degradation and Stability 25 (1) (1989) 39e47. [20] R.M. Jones, C.W. Bert, Mechanics of composite materials, Journal of Applied Mechanics 42 (3) (1975) 748. [21] Standard from the International Electrotechnical Commission [IEC].

AND

C ONNECTORS

233

[22] C.A. Charalambous, et al., Impact of photovoltaic-oriented DC stray current corrosion on large-scale solar farms’ grounding and third-party infrastructure: modeling and assessment, IEEE Transactions on Industry Applications 51 (6) (2015) 5421e5430. [23] I. Cotton, et al., Stray current control in DC mass transit systems, IEEE Transactions on Vehicular Technology 54 (2) (2005) 722e730. [24] Lokanath, Skarbek, Ramanathan and Seidel. “Considerations and structured approach for selecting and deploying climate specific polymeric wire management means” IEEE PVSEC 44.

10 Degradation Processes and Mechanisms of PV System Adhesives/Sealants and Junction Boxes Kent Whitfield Renewable Energy Technologies, Fremont, United states

10.1 Introduction Junction boxes provide an important function of providing a safe transition of power produced from the photovoltaic (PV) cells in the laminate to the external wiring system for interconnecting modules and loads. Junction boxes, also referred to as “jboxes,” are used in a wide variety of electrical applications and serve to protect the connections between wires and cables. In a PV module, however, the jbox also houses the bypass diode that provides a PV cell-protective function and it is prewired and terminated for rapid interconnection of modules in the field. The junction box enclosure consists of a body enclosure and lid both of which are almost exclusively made of injection-molded polymeric construction and are adhered to the substrate of the module with an adhesive compatible with both the junction box polymer and PV module substrate. The combination of the enclosure, bypass diode, transitioning wiring, connectors, and adhesive makes the jbox a complex component that appears deceptively simple and is not designed to be maintained or serviced. This combination of system behavior and lack of maintainability poses a reliability challenge for the PV module manufacturer, where a service lifetime of 25-years is expected. The importance of design, characterization of the failure and degradation modes, sufficient qualification and durability testing, and manufacturing quality control for the Jbox system (box and connectors) cannot be overestimated since failures of this system account for approximately 18% [1] of 180 fire incidents studied from 1995 to 2012. This study indicates that the jbox with its associated PV connector was the number one cause for a fire, with the second being inverters at 16%. The International Energy Agency’s 2014 report [2] suggests that PV connectors alone may be responsible for up to one-third of PV module fire incidents suggesting that special attention to this component is required.

This chapter describes both some of the historic failure and degradation modes and their influence on testing and certification standards as well as the new failure modes that have emerged in the contemporary jbox. Due to the frequency of PV connector and bypass diode failure, special attention is given to these components.

10.2 History All jboxes contain an electrical connection system that transfers electrical power from the laminate containing the PV cells to the wiring system that is used to deliver power to a load or inverter. For example, common electrical connection systems are composed of metallic rails with lugs, spring-clips, solder pads, or terminal barrier strips that connect the PV ribbons extending from the laminate substrate to the output cable and connector. Jboxes vary in their schemes for: bypass diode protection; water and dust Ingress Protection (IP), ranging from drip-proof to water immersion and dust-tight gasketed systems; and now a growing number includes integrated electronics such as maximum power point trackers. The majority of jboxes available today do not require, or in many cases provide, service access which is a significant departure from the historical wiring compartment jbox where the user was expected to remove the lid, remove a knock-out for a conduit system, pull wires into the box, and terminate them on lugs or terminal barrier strips (Photo 10.1). Wiring compartment-style jboxes that accommodate conduit are now rarely seen except in hazardous locations [3] Article 500, Hazardous (Classified) Locations, Classes I, II, and III, Divisions 1 and 2, where explosion-proof installation methods are required to contain a spark that might be created from damaged wire insulation from igniting a flammable gas mixture. Aside from the economic implications

Durability and Reliability of Polymers and Other Materials in Photovoltaic Modules. https://doi.org/10.1016/B978-0-12-811545-9.00010-0 Copyright © 2019 Underwriters Laboratories, Inc. Published by Elsevier Inc. All rights reserved.

235

236

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Photo 10.1 http://www.millersolar.com.

associated with conduit and wire termination in the field, this style of box contained several degradation and failure modes that could be exacerbated by field installation procedures such as:
insufficient cold impact toughness and
contact series resistance increase at the terminal due to
creep and consolidation and
contact corrosion These issues (described later) provided additional impetus for the transition to the jbox’s contemporary form. The overwhelming majority of jboxes are of polymeric construction, however, some metallic enclosures were used in the past due to their low cost and wide availability and can be seen today when thermal management is a driving consideration (Photo 10.2). The National Electrical Code [3],

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

however, proposes that the use of metallic enclosure is problematic from a balance of wiring material cost and installation time perspective. In 1984, according to Article 690 titled “Solar Photovoltaic Systems” that covers safe electrical installation practices, all exposed nonenergized metal are required to be grounded and this means that metallic enclosures require a field-implemented grounding system that adds both material and labor costs as well as installation time to the project (as witnessed by the green colored wire in Photo 10.2). These added costs run counter to the market success factors of jboxes that have largely remained unchanged since the Jet Propulsion Laboratory’s Low-Cost Solar Array Project was established in 1975 [4]. Namely, a jbox is considered marketable if the PV module manufacturer obtains the correct features and functions from the box, is able to obtain safety certification for their product, and is able to obtain the box at a volume and price considered acceptable. Often the price is considered from both an initial and cost reduced feature perspective where the cost reduction is either in the form of volume commitment and or alternative constructions. Cost reduction of a junction box or any PV module component requires a thorough understanding of failure and degradation modes to reduce the risk of premature failure. Although not meant to be a comprehensive list, the following list identifies the major degradation and failure modes considered when designing and implementing a junction box system with appropriate references to other chapters of this book.
Polymer selection for:
Cold impact toughness [2.1],
Thermal aging,
Surface voltage tracking characteristics,
Resistance to ignition,
Weatherability (UV and water exposure),
Manufacturability (mold stress and dimensional consistency) [2.3].
Adhesive selection for:
Thermal and chemical compatibility with the jbox and module substrate material [5],
Adhesive strength,
Resistance to creep at an elevated temperature [5.2],

Photo 10.2 K. Whitfield, metallic junction box.

10: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF


Flow and coverage to prevent liquid water ingress,
Electrical connection of laminate ribbons and electrical wires/cables leading to potential for:
Creep [2.2],
Contact series resistance increase (fretting, oxide and intermetallic formation) [3.1], and/or
Contact corrosion [2.3].
Bypass Diode and Blocking Diode [4]
Temperature of die and overall heat transfer rate to the environment,
Peak inverse voltage and reverse leakage current at the temperature, and
Design (hot-spot protection).

and

reverse

current

10.2.1 Cold Impact Toughness As indicated, the early style of wiring compartment jboxes used conduit to interconnect modules in the field. The requirement to remove knock-out plugs for conduit regardless of ambient air temperature posed concerns with the training and consistency of the installer performing the work. Often, the procedure consists of using a flat head screw driver placed in demarcation surrounding the knockout plug and then striking it with a hammer (Photo 10.3). This intentional impact is the reasoning behind the UL 1703, Standard for Safety Flat-Plate Photovoltaic Modules and Panels [5], Section 30 Impact Test. This test subjects critical locations on the jbox and connector to 6.8 J impacts produced from a 51 mm steel sphere falling 1.3 m to the impact site after the box and connectors have been cooled to 35 C for 3 h. This test had practical significance in the wiring compartment era where it was likely that an installer

Photo 10.3 Accessed from http://waterheatertimer. org/6x6x4.html.

PV S YSTEM A DHESIVES /S EALANTS

237

might be removing the knock-out plugs during a winter installation and it was essential that this operation produces no collateral microcracks in the compartment body that could subsequently form a liquid water ingress path and potentially result in a shock hazard or accelerated contact corrosion. Now, however, this test is less directly relevant, but often considered from the perspective of hail damage on tracking systems where junction box exposure may occur during early morning or late afternoon inclinations. The cold impact test also has a significant influence of the selection of suitable jbox and connector polymers because many have glasstransition at temperatures above 35 C.

10.2.2 Creep and Consolidation The use of terminal barrier strips or post and nut style terminations in the wiring compartment resulted in combining metals of differing thermal expansion and hardness (typically plated steel to copper). Differential expansion and contraction caused by both diurnal temperature swings and operation under load create contact pressure changes in these joints with a resulting tendency of the softer metal to creep leading to lower contact pressure and a resulting increase in series resistance. Relative motion, if present can also result in fretting that can accelerate the resistance increase. The net effect is a time and temperature-cycling-dependent loosening of wire and cable screw connections. This mechanism produces a slow and steady increase in series resistance leading in a loss of module power output and in some cases overheating of terminals leading to enclosure thermal damage or fire (Photo 10.4). The process is

Photo 10.4 Localized overheating due to cable strands becoming loose. K. Whitfield.

238

D URABILITY AND R ELIABILITY

OF

P OLYMERS

an inherently positive feedback loop with increasing series resistance resulting in larger diurnal temperature swings. To combat this performance degradation mode, some commercial PV systems adopted annual terminal and lug tightening maintenance programs that required source circuit to be taken off line, junction boxes opened, and terminals torqued to manufacturer recommendations. The ubiquitous nature of PV connectors and permanently wired jboxes has eliminated a maintenance requirement for pressure connections inside the module, but they remain relevant today in combiner boxes, disconnects, and inverters and the general guidance for maintenance of these connections remains an annual inspection and retightening to avoid performance degradation and potential safety concerns [6]. Failure of solder joints inside a jbox is also related to creep phenomena where the number of cycles to failure is dependent on the level to total plastic strain introduced into the joint as a function of differing levels of thermal expansion that may exists between the copper ribbon and electrical rail (copper alloy, plated steel, etc.) or between the bypass diode leads and rails. The function of strain relief elementsdsuch as added ribbon length and diode lead formingdsignificantly reduces the likelihood of this failure mode to be dominant within a module’s lifetime, although they may be still seen in bypass diode-to-die connections (Photo 10.5) when these components are routinely operated due to a poor installation location or as the result of a module Photo 10.5 Diode die failure from thermal cycling. K. Whitfield.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

defect and are a definite contributing factor to module performance degradation at the ribbon-to-cell soldered connection [7].

10.2.3 Contact Corrosion One of the earliest tests for evaluating the suitability of a jbox against liquid water ingress was the UL 1703 Water Spray Test where a torrential rainfall produced by three specially designed and calibrated low-pressure spray nozzles (Photo 10.6) directed a water spray at the junction box for 1 hour. Following this exposure, the product was subjected to a test of the efficacy of the electrical insulation system (dielectric voltage withstand test) and the jbox lid was opened and the interior examined for the presence of water. This test is similar to the IEC 60529, “Degrees of Protection Provided by Enclosures (IP Code)” requirements for an IPX5 rating that was present in the first edition of IEC 61730, “Photovoltaic (PV) Module Safety Qualification” standard and remains the minimum water resistance rating in the current IEC 62790, “Junction Boxes for Photovoltaic Modules e Safety Requirements and Tests.” This level of liquid water ingress protection is suitable to maintain a product against electrical shock safety risks, but may be insufficient for long-term resistance to electrical contact and terminal corrosion. The 1000 h 85 C, 85% RH Damp Heat test became a widespread test of PV cell and electrical component corrosion resistance following the

10: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

PV S YSTEM A DHESIVES /S EALANTS

239

Photo 10.6 Water spray (from UL 1703).

publication of IEC 1215 (later renumbered as IEC 61215), “Terrestrial Photovoltaic (PV) Modules e Design Qualification and Type Approval” in 1993, and later in the IEEE 1262, “Recommended Practice for Qualification of Photovoltaic (PV) Modules” in 1995. The Damp Heat test has been highly successful in identifying the potential for contact corrosion although its direct relationship to field corrosion rates is highly variable and depends on the material combination being studied [8]. Damp Heat was, in fact, the leading cause of product qualification failures as summarized by Hammond et al. in 1996 where both contact corrosion and enclosure distortion (mold stress relief) due to the use of polymers with unsuitably low softening temperatures or improper injection molding processing were found on several manufacturer’s modules. Contact corrosion is commonly seen on older wire-enclosure-style jboxes (Photo 10.7) if the pressure plate or binding screw

Photo 10.7 6/177 Solarex failures in SYSU study 2015.

contained tin, zinc, or cadmium plating since the combination of copper and these coatings can result in a corrosion-inducting galvanic potential in excess of 600 mV or more in the presence of moisture [5]. Electrolytic contact corrosion is occasionally seen in the field if liquid water is present in the jbox while the PV module is exposed to light (Photo 10.8). Photo 10.8 shows an example of such an incident and is an illustration of multimodal failure caused initially by contact corrosion. This incident was triggered when a field installation was curtailed during a rainstorm and the installation crew placed some modules in a ditch that became flooded. The jbox was not potted and although it had an o-ring, the manufacturer described the design as having an IP 65 rating (see Table 10.1) where the second digit indicates that the jbox is protected from low pressure water jets from any directions, but is not protected against immersion. The o-ring for this level of protection was formed using a thermal joining technique that was suitable for the rating, but did provide an imperfect seal against immersion due to an irregular raised surface at the joint

Photo 10.8 K. Whitfield.

240

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Table 10.1 Degrees of Ingress Protection Second Number

Test

0

No special protection.

1

Protected against vertically falling drops of water or condensation.

2

Protected against falling drops of water, if the case is disposed up to 15 degrees from vertical.

3

Protected against sprays of water from any direction, up to 60 degrees from the vertical.

4

Protected against splash water from any direction. Limited ingress permitted.

5

Protected against low pressure water jets from any direction. Limited ingress permitted.

6

Protected against high pressure water jets from any direction. Limited ingress permitted.

7

Protected against short periods of immersion in waterdbetween 15 cm and 1 m.

8

Protected against long, durable periods of immersion in water, under pressure.

Degrees of ProtectiondSecond Digit. The second digit indicates the degree of protection of the equipment inside the enclosure against the harmful entry of various forms of moisture (e.g., dripping, spraying, submersion, etc.)

(Photo 10.9). The discontinuity at the o-ring surface in combination with the slightly higher durometer at this thermal joint provided a liquid water ingress point to the box that became flooded over the course of the storm. On exposure to sun, the illuminated module generated a 35-V potential in the flooded jbox that initiated a rapid electrolytic decomposition process that was most pronounced on the positive terminals. The corrosion went unnoticed following installation and in less than 1 year several modules were discovered with warped jbox lids caused by bypass diode failure. In this situation, the module string nearest the positive output terminal had sufficient contact resistance to trigger the bypass diode into continuous operation. The bypass diode leads were connected to the electrical rails by spring contacts that had also been corroded leading to higher

Photo 10.9 K. Whitfield.

diode running temperature and eventually premature failure. It was found experimentally that a similar degree of contact corrosion could be produced in less than 1 day of water-submerged exposure to 30 V from a power supply. Although this sequence of events may seem involved and improbable, it illustrates an important concept in product degradation which is that it is a product of design (IP 65), manufacturing (o-ring joint), installation (protection from flooding), and maintenance (inspection of jboxes). In general, the effects of corrosion can be minimized through minimizing exposure of dissimilar materials, the use of water tight jboxes (IP X6 or X7), and the use of potting compounds that can minimize connection contact with liquid water.

10.3 PV Connectors Due to the expense and time required for a conduit and wire installation, the Low-Cost Solar Array Project included a program to develop a low-cost and durable module to module PV connector [9]. The multiconductor connector system described as the Cannon Sure Seal (Photo 10.10) was borrowed from the automotive industry and subjected to a test program developed by Jet Propulsion Lab (JPL) to determine the suitability for PV applications operating below 250 VDC. The test program subjected groups of samples to specific sequences of testing. The performance tests developed by JPL contained many similarities to the present IEC standard for PV connectors although as indicated in Table 10.2.

10: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

PV S YSTEM A DHESIVES /S EALANTS

241

Photo 10.10 5101-20 [9].

Table 10.2 Comparison Between the JPL’s Early Work on PV Connectors and the Present IEC Standard JPL Test Program on Low-Cost Connector for Solar-Array Modules 1977

IEC 62852:2014, Connectors for DC-Application in Photovoltaic SystemsdSafety Requirements and Tests

Dielectric withstand and insulation resistance

Dielectric voltage withstand and voltage impulse

Connector retention force

Insertion and withdrawal force



Salt spray (96 h, 35 C, and 20% NaCl)

Corrosion resistance (sulfur dioxide or IEC 60512)

Thermal cycling with humidity exposure (one cycle from 55 to þ85 C with 90% RH, approximately 1.5 h duration)

Thermal cycling (40 to þ85 C, 200 cycles)

Connector durability (5 cycles of 10 lbf for 10 min connector loading followed by disconnection and reconnection)

Bending cycling (10 or 20 N load during 90 degree flex for 100 bends)

UV exposure (102 min twin carbon-arc lamp exposure followed Weather resistance (102 min xenon exposure by 18 min water spray for 96 h total at 63 C) followed by 18 min water spray for 500 h at 65 C) Ozone exposure (72 h, 100 ppm, 50 C) NA

Ingress Protection

NA

Damp heat (1 kh at 85 C, 85% RH)

NA

Temperature rise at 85 C and full current

NA

Mechanical strength at 40 C

NA

Flammability testing (Glow wire test at 650 or 750 C)

242

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Photo 10.12 Matt Paise.

Photo 10.11 MC3 K. Whitfield.

Although the Sure Seal was reported to have met many of the requirements of a suitable PV connector, the original polymers selected for the body indicated susceptibility to UV degradation. Subsequent work by JPL [10] led to several alternative connectors from companies such as Motorola, Cannon, and Amp, however, the use of connectors remained limited for until the end of the last century. MultiContact’s MC3 PV connector system was the first widely adopted and commercially viable PV connector introduced in 1996 (Photo 10.11) [11] and provided an opportunity to limit the role of the jbox to transitioning power output from the laminate to a suitable conductor and protecting that transition and the bypass diode protection scheme. Several favorable economic impacts that led to its widespread and rapid adoption are eliminating the cost of the terminal barrier strip, reducing the mass of injected plastic since the space for internal field wiring space was no longer required and the corresponding reduction of installation time in the field. These early connectors found their way into megawatts of fielded systems and have had minimal issues, but some of the degradation modes associated with these connectors are worthy of note such as: Incomplete engagement of connectors and fretting corrosion (Photo 10.12), Use of improper crimping tools or inadequate crimping (Photo 10.13), Rodent damage (Photo 10.14), and Ground faults from wire management practices (Photo 10.15).

Photo 10.13 PVQAT TG10.

Photo 10.14 T. Zgonena

Owing to the fact that connectors may have been attributed to some 29% of field fires (based on a 2014 study of 75 incidents [2]) this component warrants special attention. The following sections expand on the four categories identified.

10: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

Photo 10.15 K. Whitfield.

10.3.1 Incomplete Connector Engagement and Fretting Corrosion PV connectors typically contain a pin that fits snugly into a barrel containing a rolled spring form to create a low resistance connection (Photo 10.16). The pin and barrel must be recessed sufficiently to remain touch safe for installers because the interconnection of modules in the field results in voltages well above the touch safe limit of 30 V for wet locations as indicated in the National Electrical Code (NEC). The requirement to be touch safe removes visual feedback from the actual electrical connection process and under various circumstances discussed later, may lead to instances where the spring is not fully

PV S YSTEM A DHESIVES /S EALANTS

243

engaged on the pin, reducing the contact area and pressure that lead to series resistance above nominal conditions (approximately 2 mU [12]). As series resistance increases, so does the temperature rise from Joule heating during operation, which creates a corresponding micromotion of the pin and barrel that is the result of thermal expansion and contraction differences between the metallic and polymeric pieces that make up the connector. The consequence is fretting where metal transfer and wear occur that may also result in oxide formation depending on the temperature rise experienced. Regardless, the outcome is a series resistance increase leading to additional temperature rise. In extreme cases, the outcome is melting of the connector housing and potentially fire or arcing (see [2]). Cold-weather installation posed a challenge for early connectors that had an interference fit rubber insulator leading to a probability of incomplete latching. The consequence often went unseen in these connectors since they did not have a locking mechanism that provided audible and visual feedback to the installer when the polymer housing fully seated. A personal experience with this problem occurred in 1999 at an installation where the ambient air temperature was below 0 C. The rubber material in question was a vulcanized ethylene propylene diene monomer (EPDM) rubber in a thermoplastic matrix of polypropylene (PP) and while the material remained well above its glass transition temperature of 60 C, the modulus had Photo 10.16 Internet grab.

244

D URABILITY AND R ELIABILITY

OF

P OLYMERS

increased to a point that it was not possible to fully seat the connector halves. The manufacturer’s remedy was to soak the parts in isopropyl alcohol to facilitate insertion, but the author decided instead to delay the installation until warmer weather prevailed. In this case a partially engaged connector was avoided, but a review of field incidents suggests that this issue remains a contemporary failure mode. The change to a locking-style connector, prompted by the 2008 National Electrical Code (NEC), sought to avoid fire and shock hazards that could develop from a partially engaged connector and from accidental disconnection while the connector was under load. Currently, the primary reasons for a lack of complete connection today are the result of quality control during the installation process, indicated to affect some 5% of installations in 2002 [13] and was more recently cited as contributing to 12% of field failures from a 2011 study of experiences with 21 module manufacturers by SunPower [14]. Another aspect of incomplete latching is the mating of connectors from different manufacturers. Although there are standard styles of connector on the market today, they are not necessarily compatible with each other due to slight variations in housing geometry leading to incomplete engagement (Photo 10.13) and possibly more concerning are degradation modes associated with different electrical connector plating materials that manufacturers select. An interesting study of baseline contact resistance changes when combining different connectors is provided by Pandey [15] where some combinations result in a more than a factor of two increase in resistance when compared with mating connectors from the same manufacturer. Even if the initial resistance is favorable, changes in plating coatings between manufacturers may accelerate wear. The predominate plating styles used on connectors are tin and silver over copper alloy parts. Tin-plating requires careful design for fretting corrosion since tin-oxide is several orders of magnitude harder than tin and forms particles that can accelerate wear. Also at elevated temperatures (above 125 C), both oxide formation and tin-copper intermetallic can form affecting both the hardness and electrical conductivity of the surface that accelerates wear [16].

10.3.2 Field Crimping Issues Connectors are evaluated for safety in conjunction with the use of a proper tool to crimp the connector to

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

the module or field wiring system. Either the connector manufacturer specifies the proper tool to use or the installation instructions include information on crimping die geometry, the number of crimps required, and similarity of crimp forces between tools [17]. Unfortunately, field installation instructions are occasionally not followed and the consequence can lead to field failure of the connector (Photo 10.13). This is a quality issue during installation, and as such, the timeframe by which the failure of the component occurs is often within 5 years with the highest frequency of events occurring within the first year of operation [1]. The study by Laukamp that analyzed 180 cases of PV system-related fires that occurred between 1995 and 2012, of which 110 could be attributable to a root cause, indicates that approximately 37% could be attributed to installation faults with connectors representing 8% of the total number of incidents and cross-combining connectors (see previous section) and improper crimping topping the list of root causes. Proper crimping tools are especially important to properly consolidate and compress uniformly around fine-stranded cables preferred in PV systems for their increased flexibility. Without uniform crimping, fine wire strands may become loose leading to increased resistance through reduced conductor ampacity leading to additional heating [18].

10.3.3 Rodent and Insect Damage The latching-style of connector was often a source of rodent damage (Photo 10.14); however, all flexible cords and cables have been subject to attack from rats and squirrels. This source of damage is more problematic on residential applications due to the close proximity of the PV module to the roof surface that prevents visual inspection of the damage and, due to potential debris accumulation, may worsen the outcome of an arcing event beneath the module. The solution to this problem is primarily through wire management practices and specifically to either remove cables and connectors from the path of a rodent or guard the installation from access. The issue of ground faults, shock hazard, and fire potential due to this form of damage has become prominent enough to warrant changes in Ontario’s 2016 Electrical Safety Code [19] where guarding against access is now a code requirement for PV system installation.

10: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

10.3.4 Ground Faults Ground faults are the result of a breach of the electrical insulation system allowing, under the right circumstances, current to pass from the current carrying conductors through to the grounding electrode system typically via equipment ground conductors or grounded frames. Ground faults can be particularly troublesome to locate on PV systems due to the long cable runs used on commercial and industrial PV systems and the variety of grounded metal equipment used for support, enclosures, and conduit. In the case of PV connectors, ground faults represent not only an energy hazard (which is typical of equipment and wiring ground faults), but a shock hazard as the connector is typically exposed and its relationship to a grounded element is not fixed meaning accidental handling of a connector experiencing a ground fault could alter the fault current path through the handler. Ground faults often are the result of a failure of the electrical insulation system and exposure to liquid water forming a path from the electrically live parts to ground. Three of the more commonly found connector related ground faults derive from manufacturer quality control of the cable gland securement (Photo 10.17), incomplete connector engagement

Photo 10.17 K. Whitfield.

PV S YSTEM A DHESIVES /S EALANTS

245

(previously described), and wire management practices in the field. Cable gland securement requires the uniform compression of a polymer ferrule against the cable and the connector body. If the manufacturing facility does not have appropriate controls, variation in the torque of the cable gland nut can result in direct water ingress around the ferrule, or can damage the cable insulation with a potential water ingress potential at the conductors. Installations that place the PV connector body at the lowest position in the module-to-module connection, or in exposed cable trays may be at a higher risk from water exposure (Photo 10.18). Incomplete connector engagement poses similar risk for water ingress and has been previously discussed; however, other failure mechanisms such as disconnection under load and series resistance increase causing thermal failure to dominate this particular installation error. The last category of wire management practice is a case where an effort to improve cosmetics of an installation inadvertently can lead to a field failure. It is often the case that module manufacturers design their products for installation in both portrait and landscape orientations and due to the jbox location, one style will dictate the length of the cable required to perform a module-to-module interconnection without placing strain on the cables or jbox. The result is that the orthogonal orientation typically has an excess of cable that should be constrained. As shown in Photo 10.19, the wire management strategy used in this installation involved tightly coiling the excess cable between modules and to secure the coil to a cable tray. The radius of the coil was small enough that a gap opened in the connector body

Photo 10.18 K. Whitfield.

246

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Photo 10.21 K. Whitfield. Photo 10.19 K. Whitfield.

allowing liquid water to enter and its proximity to the grounded metal tray established a high resistance ground fault that over time lead to copper corrosion that (Photo 10.20) lowered this fault resistance. The eventual failure mode was an arc fault that caused fire damage to both the module and roof. During investigation of the incident, it was possible to isolate wire management practices by studying the same module at different sites (similar climates) where evidence of copper corrosion around the connector body to cable interface was absent when larger radius coiling (or no coiling) had been used. It was also possible to replicate this failure mode in a laboratory setting at a shorter time scale by replicating the coil radius and increasing the applied voltage between the cable tray and the connector cable (Photo 10.21). Rigid connector bodies are not immune from wire management-induced failures since cable bending strain will produce a non-uniform stress on the polymer ferrule that may lead over time to plastic flow of the ferrule and a potential water ingress path.

Photo 10.20 K. Whitfield.

10.4 Bypass Diodes The function of a bypass diode in a photovoltaic (PV) module is to limit the amount of current flowing through a reverse-biased cell. PV cell reverse biasing is a condition that can occur whenever a cell exhibits reduced operating current with respect to other cells in the same circuit. If the cell is part of a much larger string of cells that are operating at a higher current level, the reduced current cell can operate at the larger current only through reverse bias operation. The most common reason for a cell’s reduced current operating condition is localized shading (e.g., debris, soiling, or shadowing obstruction) on the superstrate. Other conditions that may cause reduced current operation include extremely mismatched cell electrical characteristics in the string, cell damage (e.g., cracks) that remove a portion of cell area from electrical current production, or in concentrator PV, local cell misalignment with respect to the sun. The need for bypass diodes has been well documented and accepted as an appropriate protective strategy against so called “hot spot” operation. From a 1993 study of 20,000 modules that were fielded prior to the Jet Propulsion Labs’ Block V requirements, 10,000 had ceased power production and 90% of these were stated to have failed for common reasons including “. hot spot formation from inadequate or lack of bypass diode protection” [20]. This is because a reversed bias cell dissipates heat proportional to the degree of reverse bias voltage and the current flowing through the cell. This heating can be uniform across the cell surface, or can be extremely localized in the form of a weakened dielectric breakdown site referred to as “current shunts” that are similar to the breakdown behavior exhibited by standard diodes when they are subjected

10: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

Photo 10.22 K. Whitfield.

to large reverse bias compared to their breakdown voltage. Hot spot temperatures can be extreme enough to melt through polymer backsheets, shatter glass, and ignite flammable materials in contact with the hot spot. A 1997 study indicated that temperatures exceeding 150 C were reasonable for crystalline silicon modules [21], however, as shown in Photo 10.22, much higher temperatures can occur as in this test where the diodes were removed from a module to understand the consequence of a bypass diode failing in an open-circuit condition. The level of required bypass diode protection is determined based on the cell’s reverse breakdown voltage and operating temperature. Breakdown voltage is a continuous random variable having a central tendency and a range often associated with a level of localized defects within the cell that create localized dielectric weak spots. Temperature has an overarching effect of modulating the reverse leakage current level at a given level of reverse voltage; the higher the operating temperature, the higher the reverse leakage current level. Given the vagaries associated with confirming that a PV module was designed with a suitable level of bypass diode protection, a standard performance test was constructed to provide evidence of satisfactory hot spot protection. The current Hot Spot Endurance tests (IEC 61215 clause 10.9 or UL 1703’s Section 39) have a common antecedent in the JPL Block V hot spot endurance test published in 1981 [22] that was described as a “complicated 100-h[our] cyclic procedure” that included test cell selection based on dark current-voltage characteristics and required the use of infrared lamps for simulating operating

PV S YSTEM A DHESIVES /S EALANTS

247

temperature, visible light exposure to increase current flow through the cell, and external power supplies to drive the cell into a reverse voltage condition. The hot spot endurance tests fulfill a role of determining if a PV module will suffer a degrading hot spot due to inadequate bypass diode protection, but necessarily assumes that the diodes are functional. As indicated by Smokler [23], “Since diodes are required to perform their function occasionally or continuously over the life of the array, an important consideration is their long-term reliability. Important parameters influencing by-pass diode reliability include derating of the diode characteristics, adequacy of the heat-sink design, and the expected worst-case field thermal environments.” General design criteria are stated in this document as shown in Table 10.3: Based on these guidelines and the PV cell characteristics at that time, the general rule of thumb was to use standard Si p-n diodes at a rate of 1 per 18 series-connected cells. Currently, due to improvements in crystalline cell manufacturing that has increased breakdown voltage and the prevalent use of Schottky bypass diodes with their lower conduction voltage (discussed later), it is now commonplace that one bypass diode protects 24-series connected cells. Concurrently the average current of a module has approximately doubled since the 1990s due to increases in wafer diameter. Both changes result in more than 250% increase in power available to a hot spot placing even more importance on the bypass diode’s reliability to prevent hot spot field failure of the PV module.

10.4.1 Diode Evaluation There are two primary ways that have been used to confirm that a photovoltaic module has adequate bypass diode protection: 1. examining performance under simulated shading conditions designed to bring about approximately worst-case hot-spot conditions and 2. examining the thermal performance of the bypass diode when passing 100%e125% of short-circuit in an elevated ambient chamber. Item 1, the Hot Spot Endurance Test (MQT 09 in the IEC 61215e2:2016, “Terrestrial Photovoltaic (PV) Modules e Design Qualification and Type

248

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Table 10.3 JPL Recommendations for Diode Reliability Maximum Allowable Junction Temperature

Derated Temperature for LongTerm Reliability

p-n

175 C

125 C

Schottky

125 C

Diode Type

Applicable Field Conditions

Project

75 C

100 mW/cm 40 C 1.5*Isc

2

100 mW/cm2 40 C 1.0*Isc Reference/ Illustration

Problem

Root Cause (per Diode Manufacturer)

IEC61215 and UL1703 Certifications

Diode Manufacturer “A” Failure after TC200 (shorted condition)

ESD Damage

IEC61215 and UL1703 Certifications

Diode Manufacturer “B” Diode Failure after TC50 (shorted condition)

ESD Damage

IEC61215 and UL1703 Certifications

Diode Manufacturer “B” Failure after Mechanical Load, Hail Impact, Hot Spot Endurance (shorted condition)

ESD Damage

Approval” version) is a multistep evaluation that determines if the module design has adequate bypass diode protection. It is not an evaluation of the diode, but an assessment of the hot spot behavior of the module and specifically specially selected cells presuming that the diode is working perfectly. This test begins by categorizing the electrical arrangement of interconnected cells being protected by a bypass diode into series, parallel-series, or series-parallel to clarify the test method to be used in a worst-case hot spot condition. Next, cells are evaluated for their reverse-bias leakage-current characteristics from which cells are selected based on the electrical arrangement for the module. Finally, a shading condition is selected to maximize a hot spot effect and this condition is held from 1 to 5 h depending on the cell temperature stability. Provided there is no evidence of thermal failure and the module remains electrically functional, the bypass diode strategy employed is considered sufficient.

Item 2, the Bypass Diode Thermal Test (MQT 18.1 in the 2016 edition) determines the adequacy of the thermal management of the diode in the jbox when the diode is in a conduction state and in an elevated air temperature of 75 C and operated at module short-circuit conditions for 1 h followed by 125% of module short-circuit for 1 h. The diode is to not exceed the rated maximum junction temperature, remain functional, and create no evidence of thermal damage to the jbox in this test.

10.4.2 Perceived Gaps and PVQAT TG4 In 2011, an international task group was formed to study bypass diode testing adequacy as part of a larger initiative called the PV Quality Assurance Task Force (PVQAT). Task Group 4, Diodes, Shading, and Reverse Bias, was specifically charged with identifying gaps in the evaluation of these

10: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

components and found that qualities of bypass diode not characterized by the product standards included: 1. Accuracy of the diode manufacturer’s technical data sheet, specifically: a. maximum junction operating temperature when indirectly measured from the case or lead, b. forward current rating at elevated temperatures, and c. reverse voltage rating at the maximum ambient temperature. 2. Assessment of a diode’s potential for thermal runaway failure in a specific jbox design, and 3. Electrostatic discharge susceptibility.

10.4.2.1 Junction Temperature In 1984, Underwriters Laboratories, working under a joint contract with the Jet Propulsion Lab and the Department of Energy, indicated that the proposed Standard for Safety, UL 1703, contained no minimum requirements for bypass diodes, but that a component certification program could be created to verify the manufacturer’s technical data sheet information [24]. The component verification program was not created and both the IEC performance (61215) and safety (61730) standards have relied upon the diode manufacturer’s technical datasheet for key operating parameters identified in 1 (above) that predominantly rely on the diode not exceeding its maximum junction temperature. While the temperature limit of the diode should remain the providence of the manufacturer, an issue has been how the junction temperature is estimated from the experiment. Normally, a thermocouple is applied to the exterior of the diode case or lead and then the manufacturer’s thermal resistance parameters are used to convert the measurement to a junction temperature. Slight variations in the manner of adhering the thermocouple and the thermocouple’s influence on heat transfer from the diode both can easily lead to large differences in inferred junction temperature. Based on work done both in PVQAT and the Working Group for IEC 61215, the 2016 version of this standard has been changed to include a characterization procedure that holds the diode at four different steady state temperatures from 30 to 90 C and introduces a 1 ms pulse of full load current while measuring the

PV S YSTEM A DHESIVES /S EALANTS

249

resulting voltage drop across the diode. The data are placed into a graphical form and can be interpolated or extrapolated to determine junction temperature as a function of current and local ambient air temperature. This method does away with the inaccuracy of the previous temperature and conversion measurement process.

10.4.2.2 Thermal Runaway A mode of failure that was not assessed in IEC 61215 standard is that of thermal runaway. Photo 10.23 shows a poorly located PV system that resulted in daily shadowed operation, requiring daily bypass diode operation. In this installation, the bypass diodes failed and the consequence was a fire that originated from the jbox, possibly due to reverse thermal runaway. In a routinely shaded system, the diode will transition from shaded cell operation under load, with a resulting significant increase in junction temperature, to a reversed biased condition when the shadow passes. As reverse bias leakage current is directly proportional to temperature a critical temperature exists (Photo 10.24) that if passed, will enable sufficient reverse leakage to continue passing through the diode resulting in further temperature rise, a positive feedback situation that does not end until failure has occurred. This is referred to as a thermal runaway condition and a new test method has been published by IEC TC 82 Working Group 2 to assess a design for this failure mode. IEC 62979 Photovoltaic ModuleseBypass DiodeeThermal Runaway Test subjects the diode in the jbox to forward current of 125% of module Isc while the module is held at 90 C in a chamber (this

Photo 10.23 UL.

250

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Photo 10.24 2014 NREL PVMRW (online through OSTI.gov).

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Thermal Runaway Diode Vf x If Diode Vr x Ir Cooling Curve - Low Thermal Resistance Cooling Curve - High Thermal Resistance

A Power

Critical Temperature B

No Thermal Runaway

TAmbient

temperature may be lowered to 75 C for modules intended for open-rack installation only). The test condition is then switched to reverse bias the diode at the module’s string open-circuit voltage within a 10 ms period and the diode temperature monitored for several seconds to determine if thermal runaway occurs. The critical aspects of the test are both the diode’s maximum junction temperature, maximum reverse voltage, and the thermal management within the junction box. The same diode type may pass in a jbox with adequate heat sinking and fail in another jbox design despite being tested to the exact same current and reverse voltage test conditions [25].

10.4.2.3 Other Qualification Tests Bypass diodes are part of a product qualification testing procedure that includes accelerated tests such as: 1. 200 thermal cycles from 40 to þ85 C with maximum power current flowing through the module whenever its temperature is transitioning from 40 C to 80 C. From the bypass diode’s perspective, this test places the diode into a reverse-bias condition whenever maximum power current is flowing, 2. 1000 h at 85 C and 85% relative humidity (RH), however, this test does not include a voltage bias component so is strictly a test of the packaging resistance to moisture intrusion or inherent corrosion resistance of the diode. 3. 50 thermal cycles from 40 to þ85 C and 10 humidity-freeze cycles from 20 h at 85 C and

Junction Temperature (TJ)

85% RH to 40 C and back to high temperature and humidity. These tests also do not include a voltage bias component. Electrical performance of the PV module, following these stress tests, will catch a bypass diode that fails in a shorted condition (vast majority of failure are in a shorted condition). A possible criticism, however, is that the tests designed specifically to address the suitability of the PV module’s bypass diode strategy are very short duration tests such as the approximately 2-h long Bypass Diode Thermal Test and the approximately 5-h long Hot Spot Endurance tests. A paper published in 2008 [26] showed a dramatic increase in the failure rate of bypass diodes and indicated it as the number one failure mode for crystalline silicon modules for the period of time from 2005 to 2007. As noted in the article, “The postperformance failures in the thermal-cycling and ultraviolet chamber tests are significantly influenced by the shorting problems related to the bypass diodes. The shorting problems of the bypass diodes appear to be associated with the prolonged thermal stresses of the diodes in the chambers.” During this time, some diodes experiencing similar failures in qualification testing were sent for failure analysis that suggested an electrostatic discharge event as the likely cause (see Table 10.3). In all cases, the autopsy of the diode die itself revealed a localized short-circuit created at the perimeter of the die, no damage to the packaging and no signs of overheating stress on any adjacent materials. These symptoms are not consistent with overheating, but cannot alone rule out an

10: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

overheating failure. In an overheating event, typically, there is a failure of the packaging material, which is often plastic or epoxy, or signs of thermal degradation for material in contact with the diode case. In extreme cases, plastic packages will rupture leaving visible out gassing signs on the surrounding material. Examples are shown in Photo 10.25. As the current product standards do not assess diode surge susceptibility, it is reasonable to ask whether or not products may be unintentionally being subjected to surges which produce a failure that is attributed to other causes.

Photo 10.25 K. Whitfield.

PV S YSTEM A DHESIVES /S EALANTS

251

10.4.2.4 Electrostatic Discharge Susceptibility Electrostatic discharge (ESD) is a wellcharacterized failure mode in the semiconductor industry and is routinely mitigated through a combination of design considerations (e.g., wire routing, electromagnetic filters, and enclosure protection), handling procedures (e.g., wrist straps and conductive mats), and active static control equipment (e.g., brushes, ion generators, and humidity control). Generally, the design, procedures, and active equipment remain in place until the product is protected to a suitable degree that standard handling procedures may be employed without incurring failures due to ESD. This is not necessarily the case in the PV industry. Currently, there are a wide range of ESD protection strategies in place at manufacturing facilities, but there is a commonality in that once the product leaves the manufacturing floor there is little to no intrinsic ESD protection provided with the product. This means that the electrical connectors are not intrinsically safe from a surge event and there are no transient voltage suppressors normally integrated into PV modules. Furthermore, many certified body testing laboratories struggle with attempting to coordinate environmental chamber tests among multiple clients, each potentially having their own connection means. As a result, the connectors are often simply cut off of the module and the cable is stripped back so that the bare conductor can be connected to power supplies, flash solar simulators, etc. In doing so, it is possible for the PV module electrical circuit to be subject to unintentional ESD from either charged devices or through human contact. The importance of this is while Schottky diodes are advantageous from having a low operational voltage drop as compared to a p-n junction diode, they are metalesemiconductor diodes having lower reverse voltage ratings and therefore higher intrinsic susceptibility to voltage breakdown behavior that may come from ESD. To assess the susceptibility of a selected bypass diode to ESD-induced failure, PVQAT Task Group 4 initiated work to develop a method for determining the surge voltage limit of bypass diodes. As described in IEC TS 62916, Photovoltaic ModulesdBypass Diode Electrostatic Discharge Susceptibility Testing, the method consists of a step-stress test to failure program with 10 diodes surge tested from 5 to 30 kV in 5 kV steps. A handheld multimeter with a diode check function that

252

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

1. changes in adhesive selection without proper qualification testing, 2. failures of qualification testing to address failure modes associated with adhesive creep, and 3. failure of adhered surfaces due to inadequate quality control.

Photo 10.26 K. Whitfield.

limits the available current to the low milliamp range is used to determine between stress steps if failure has occurred. Samples are tested with a positive surge directed at the cathode-side with a 10 second thermal relaxation time between surges. After each set of 10 surges, diodes are checked and failures tallied until the end of the program. The failing data are then plotted on a Weibull scale (Photo 10.26) which allows a continuous distribution to be fit through the points and this enables one to predict the failure rate that should occur in an environment having a predetermined level of ESD stress. At a minimum, a manufacturer is able to identify potentially weak ESD diodes from robust ones even though they have the same maximum junction temperature and current and voltage ratings. Such testing may also suggest a need to implement a ESD hygiene program in a manufacturing facility such as ANSI/ESD S20.20 Electrostatic Discharge Certification and as recommended in IEC TS 62941:2016 Terrestrial Photovoltaic (PV) ModulesdGuideline for Increased Confidence in PV Module Design Qualification and Type Approval.

10.5 Junction Box Adhesion Junction box adhesion is not a frequent field failure mechanism, but has occurred sporadically as the result of:

The majority of adhesives used on PV modules are one or two-part silicones due to their excellent flexibility, electrical insulation, resistance to high temperature, and primerless adhesion qualities. The one-part silicone systems are moisture curing and contain either neutral oxime or alkoxy cross-linking packages that do not trigger a corrosion reaction with metal that may occur with acetoxy that produces acetic acid. The main issue with one-part systems is that green strength is often not achieved for several hours requiring manufacturing lines to set aside curing space that might otherwise be used for additional manufacturing capacity. This has led to a rise in two-part curing systems, a class of hot-melt adhesives and the use of adhesive tapes that can adhere the junction box to the substrate with minimal to zero cure time and therefore, better space utilization.

10.5.1 Adhesive Change Compatibility between the PV module substrate, adhesive, and jbox material is often demonstrated using the Retention of Junction Box Test (MQT 14.1, IEC 61215:2016) or Robustness of Terminations Test (Clause 14.1, IEC 61215:2005) that follows a sequence of preconditioning tests that are designed to bring about weakening of adhesive bonds. This sequence consists of a 15 kWh/m2, low-dosage UV preconditioning test at 60 C that is intended to damage susceptible and exposed adhesive bonds. This is followed by 50 thermal cycles from 40 to þ85 C that uses differential thermal expansion mismatch between materials to initiate a crack or tear in adhesion interfaces. This is followed by 10 humidity-freeze cycles where the module is held at 85 C and 85% RH for 20 h followed by a rapid plunge to 40 C. During the transition from humidity to freeze, the module passes through a dew point condition resulting in the formation of liquid water on surfaces and in any formed cracks that upon freezing that, due to ice expansion, wedge-apart interfaces that have been damaged. The Retention of Junction Box test then subjects the jbox to a 40 N force that will fail marginal designs. Those that do

10: D EGRADATION P ROCESSES

AND

M ECHANISMS

OF

not mechanically fail, but have formed minor separations that form a continuous liquid water path to electrically live parts are often discovered following a Wet Leakage Current Test (MQT 15, IEC 61215:2016, or clause 10.15, IEC 61215:2005). This test partially submerges, or fully wets exposed surface of the jbox with a conductive solution and measures leakage current from current carrying parts inside the module to the conductive solution at the maximum system voltage (typically 1000 V to 1500 VDC). Visually imperceptible leakage paths often result in failure with this test. The USA requirements for PV module safety have a somewhat similar approach to the concerns of adhesive failure of jboxes that involves a fewer type of stress exposure. Specifically, UL 1703 conducts 200 Thermal Cycles (Section 35) from 40 to þ90 C followed by a pull test (Wiring Compartment Securement, Section 42) at 155 N, which is almost 300% of the IEC equivalent force, and 10 humidity freeze cycles (Humidity Test, Section 36) on different samples followed by a Wiring Compartment Securement test. These modules also undergo a Wet Insulation Resistance Test (Section 27) that is similar to the IEC test. The major difference from the IECstyle investigation is that the UL thermal cycle and humidity freeze tests are performed individually, not in sequence, and there is no UV preconditioning at the start. Both the IEC and UL styles of adhesion determination have been effective at uncovering incompatible substrateeadhesive and jbox systems although material compatibility failures were also found in the Wet Leakage Current test following Damp Heat exposure (1000 h at 85 C and 85% RH [27]) tied to deadhesion. As such the combination of humidity freeze, thermal cycle, damp heat, and a mechanical pull test and wet leakage current are typical qualification requirements that should be considered by the manufacturer when changing an adhesive material. Photo 10.27 is an example of an adhesive change with a resulting failure.

10.5.2 Adhesive Creep A class of hot melt adhesives was introduced in 2009 that were employed on some modules for complete junction box adhesion with a consequence that in elevated ambient air temperature climates, some junction boxes experienced creep. These adhesives are also used in manufacturing to provide fast

PV S YSTEM A DHESIVES /S EALANTS

253

Photo 10.27 K. Whitfield.

“green” strength while a secondary adhesive, such as a room temperature vulcanizing (RTV) silicone cured. The standard qualification sequences do not address a constant force during the stress tests (such as thermal cycle, damp heat, or humidity-freeze) and this gap is the reason that a junction box creep test was developed [28] and is evaluated without added load in the Materials Creep Test (MST 37, IEC 61730e2:2016) where the module is held vertically at 105 C for 200 h to look for signs of movement. Modifications to the module qualification standards are planned to add load to the Materials Creep Test to address this important failure mode since it has a direct electrical shock consequence.

References [1] H. Laukamp, G. Bopp, R. Grab, C. Wittwer, H. Haberlin, B. van Heeckerem, et al., PV fire hazard - analysis and assessment of fire incidents, in: Proceedings of the 28th EU PVSEC, vol. 8, 2013. [2] M. Kontges, S. Kurtz, K. Berger, K. Kato, T. Friesen, H. Liu, M. Van Iseghem, Performance and Reliability of Photovoltaic Systems, Subtask 3.2: Review of Failures of Photovoltaic Modules, International Energy Agency, 2014. [3] NEC, 2, National Electrical Code. NFPA 70. National Fire Protection Association, 2017. [4] B.H. Associates, Low-Cost Solar Array Project, Engineering Area, Jet Propulsion Laboratory, Pasadena, 1979. [5] U.L. Inc, Standard for Flat-plate Photovoltaic Modules and Panels, 2017. UL 1703.

254

D URABILITY AND R ELIABILITY

OF

P OLYMERS

[6] N.A. Group, Best Practices in Photovoltaic System Operations and Maintenance, second ed., National Renewable Energy Laboratory, Golden, 2016. [7] N. Bosco, T. Silverman, S. Kurtz, Climate specific thermomechanical fatigue of flat plate photovoltaic module solder joints, Microelectronics Reliability (July 2016) 124e129. [8] M. Kempe, J. Wohlgemuth, Evaluation of temperature and humidity on PV module component degradation, in: IEEE 39th Photovoltaic Specialists Conference (PVSC), Institute of Electrical and Electronics Engineers (IEEE), Tampa, 2013, pp. 0120e0125. [9] A.H. Cantu, Test Program on Low-cost Connector for Solar-array Modules, Jet Propulsion Laboratory, Pasadena, 1977. [10] R. Ross, M. Smokler, Flat-plate Solar Array Project Final Report, Jet Propulsion Laboratory, Pasadena, 1986. [11] C. Podewils, R. Dupont, M. Hirsch, Fresh connections, Photon International (2010) 54e56. [12] B. Yang, R. Sorensen, P. Burton, J. Taylor, A. Kilgo, D. Robinson, J. Granata, Reliability model development for photovoltaic connector lifetime prediction capabilities, in: IEEE 39th Photovoltaic Specialists Conference (PVSC), IEEE, Tampa, 2013. [13] P. Vanbuggenhout, M. Rekinger, G. Masson, T. Tsoutsos, Z. Gkouskos, S. Tournaki, PVTRIN: Training of Photovoltaic Installers, Catalogue of Common Failures and Improper Practices, EU: European Photovoltaic Industry Association, 2011. [14] D. Degraaf, R. Lacerda, Z. Campeau, Degradation mechanisms in Si module technologies observed in the field; their analysis and statistics, in: PV Module Reliability Workshop, NREL, Golden, 2011, pp. 9e25. [15] S. Pandey, Evaluation of various PV module cable connectors and analysis of their compatibility, International Journal of Current Engineering and Technology (2017) 1721e1727. [16] Y.W. Park, T.S. Narayanan, K.Y. Lee, Fretting corrosion of tin-plated contacts, Tribology International (2008) 616e628. [17] U.L. Inc, UL 6702, Standard for Connectors for Use in Photovoltaic Systems, Underwriters Laboratories, Inc, Northbrook, 2017.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

[18] NEMA, Bulletin No. 105, Fine-Stranded Cable Conductors, National Electrical Manufacturers Association, Arlington, 2012. [19] O.E. Code, Bulletin 64-4-0, Wiring Methods for Solar Photovoltaic Systems, Electrical Safety Authority, Mississauga, 2016. [20] A. Rosenthal, M. Thomas, S. Durand, A ten year review of performance of photovoltaic systems, in: 23rd IEEE Photovoltaic Specialists Conference, IEEE, Louisville, 1993. [21] W. Herrmann, W. Weisner, W. VaaBen, Hot spot investigations on PV modules e new concepts for a test standard and consequences for module design with respect to bypass diodes, in: 26th IEEE Photovoltaic Specialists Conference, IEEE, Anaheim, 1997. [22] C. Osterwald, T. McMahon, History of accelerated and qualification testing of terrestrial photovoltaic modules: a literature review, Progress in Photovoltaics: Research and Applications (2008) 11e33. [23] M. Smokler, D. Otth, R. Ross Jr., The Block program approach to photovoltaic module development, in: 18th IEEE PV Specialists Meeting, IEEE, Las Vegas, 1985. [24] U.L. Inc, Safety-related Requirements for Photovoltaic Modules and Arrays, Final Report No. DOE/JPL 955392-2, Underwriters Laboratories Inc, Northbrook, 1984. [25] N. Shiradkar, E. Schneller, N. Dhere, V. Gade, Predicting thermal runaway in bypass diodes in photovoltaic modules, in: IEEE 40th Photovoltaic Specialist Conference, IEEE, Denver, 2014, pp. 3585e3588. [26] G. TamizhMani, T. Arends, J. Kuitche, B. Raghuraman, W. Shishler, K. Farnsworth, et al., Failure analysis of desing qualifications testing: 2007 vs. 2005, Photovoltaics International (2008) 112e116. [27] G. TamizhMani, Experience with qualification and safety testing of photovoltaic modules, in: PV Module Reliability Workshop, NREL, Golden, 2010, pp. 332e335. [28] D. Miller, J. Wohlgemuth, Examination of a junction-box adhesion test for use in photovoltaic module qualification, in: Proceedings SPIE 8472, Reliability of Photovoltaic Cells, Modules, Components, and Systems V, SPIE, San Diego, 2012.

11 Accelerated Environmental Chambers and Testing of PV Modules Sean Fowler Q-Lab Corporation, Westlake, Ohio, United States

11.1 Introduction Photovoltaic (PV) modules are unique electrotechnical products. They must be exposed to sunlight to function, and shielding them from other elements of the weather detracts from their primary purpose. Nearly all other electronic devices are designed to keep out humidity and survive heat by performing their primary function within an enclosure, and durability testing of these products reflects this focus. Environmental testing focuses on these factors, and many associate the term “environmental chamber” with devices that control these two parameters. These testing devices are designed and built with varying capabilities to control hot and cold extremes and cycle between conditions, which is a subject of this chapter. According to the International Electrotechnical Commission (IEC) Technical Committee 82 (TC82), which oversees standards for PV modules, an environmental test is one “in which a product is exposed to simulated environmental conditions such as temperature, wind, rain, snow, hail, or humidity.” The tests in the IEC module qualification standards address these conditions as well as mechanical loading, both static and dynamic. In other fields concerning electronic devices, IEC TC104 (TC104) on environmental classification and testing describes these stresses and adds many others, including dust ingress, abrasion due to wind-blown sand and dust, seismic shock, mold growth, vibration from mechanical systems, and others. Two other stresses included in TC82’s list of environmental tests are ultraviolet light and salt mist exposures. Their omission from the committee’s definition of environmental tests implies that they exist apart from environmental testing or as an afterthought, despite the necessary exposure of PV modules to sunlight and in some cases corrosive conditions. It is true that weatheringdthe field that

encompasses ultraviolet light exposuresdand corrosion testing have existed as separate entities from environmental or climatic testing, and each has a long history with its own technical standards. Although currently there is no standard definition of weathering, the one proposed is “photo-induced changes resulting from exposure to the radiant energy present in sunlight in combination with heat (including temperature cycling) and water in its various states, predominately as humidity, dew, and rain.” The salt mist test is a key component of the field of atmospheric corrosion testing. A concise definition of atmospheric corrosion is difficult to formulate, but one can be stated in two parts. “Corrosion is an electrochemical process that returns refined metals to their more natural oxide states; atmospheric corrosion is a degradation process that takes place in a film of moisture on a metal surface, where the film may be so thin that it is invisible to the naked eye.” Weathering tests have not been strongly emphasized in the history of PV module testing, partly because of the general emphasis on heat and moisture in the testing of electrotechnical devices. Another reason is the difficulty in creating multifactor environments that expose large specimens to simulated sunlight in combination with heat and moisture. Typically, chambers that attempt to do this must sacrifice some facet of performance. They may sacrifice accuracy in their simulation of the UV portion of sunlight in some cases. In others, they lack moisture control or the capability to create significant thermal cycling within the exposure. Cost is also a factor. Complex chambers are expensive to design, build, and maintain, reducing the volume of data any organization can obtain within a practical testing budget. In turn, this reduces the statistical significance of the data coming from them. An alternative, economical approach is needed. In recent years, the PV industry has started to adopt some practices from other industries, such as

Durability and Reliability of Polymers and Other Materials in Photovoltaic Modules. https://doi.org/10.1016/B978-0-12-811545-9.00011-2 Copyright © 2019 Sean Fowler. Published by Elsevier Inc. All rights reserved.

257

258

D URABILITY AND R ELIABILITY

OF

P OLYMERS

automotive and building materials. In 2017, the PV industry accepted its first international standard for testing and qualifying polymeric materials for use in modules, IEC 62788-7-2. These tests can be performed on much smaller specimens, allowing the use of standardized weathering test chambers. This development is akin to qualification of automotive paints, which are tested independently according to standard methodologies specified by the car and truck manufacturers. The principle is that it is not necessary to test the entire car in order to qualify the paint’s durability. These tests have been inserted into the international qualification regime to supplement the temperature- and humidity-focused tests that have been in use. This chapter will discuss the chambers used for environmental testing of PV modules and their components. The focus will be on describing the technology behind the environmental stress tests used by the industry. Special consideration will be given to standard weathering chambers due to their recent adoption by IEC TC82 in its module qualification regime. Perhaps a better understanding of the tools will improve how test methods are both conceived and carried out. The terms “temperature and humidity chamber,” “environmental chamber,” and “climatic chamber” will be used interchangeably. Weathering and corrosion chambers will be considered as special cases of climatic chambers because they incorporate these basic stresses while introducing additional ones. Sunlight simulation and salt spray add complexity to climatic tests and create unique challenges controlling the parameters taken for granted in traditional climatic chamber testing.

11.2 The Basics of Temperature and Humidity Chambers Most climatic chambers are cuboid shaped, and they are typically specified or categorized by the volume inside the exposure area. Small, benchtop chambers may have capacity in the tens of liters, while large chambers can be sized to test an entire car or truck. Beyond the size, chambers are characterized by the minimum and maximum temperature and relative humidity levels achieved and the rate of change between controlled conditions. Most chambers are constructed with insulated walls to keep external surface temperatures at a safe operating

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

level, maximize performance and energy efficiency, and prevent condensation formation on interior chamber surfaces. The amount of insulation required depends on the high and low temperature extremes of the chamber operating specifications. Controlling temperature and relative humidity requires the ability to add and remove heat and moisture from the air. Although this seems obvious, and the essential technologies for achieving these capabilities are old, the systems required to achieve test conditions required by many applications can be surprisingly complex. The next sections will discuss these capabilities in relation to common environmental testing requirements for PV modules.

11.2.1 Air Circulation Air movement within a chamber is controlled by one or more blowers, and the system typically includes a damper system which allows fresh air to mix with exhaust air as needed to achieve programmed conditions. Damper systems vary in complexity. Simple systems may include a single damper between air intake and exhaust to control the ratio of fresh and recirculated air within the system. Conditioning processesdheating, cooling, drying, or moisturizingdmay be applied to the fresh air prior to mixing with exhaust air or after mixing occurs. More complex systems may include multiple blowers and dampers to combine air streams from different conditioning processes for precise control. These streams could be dry and water-saturated air combined in proportion to the relative humidity set point or hot and cold air streams similarly mixed. Regardless of the chamber’s complexity, air flow into and out of the system is equalized to avoid pressurizing or depressurizing the test space, with a minor exception involving a particular dehumidification technique which will be discussed later. When maintaining steady state conditions, relatively little air flow through the test space is required provided the chamber is sufficiently insulated and the devices under test (DUT) are not significant heat contributors. Higher air flow is normally required during cycling between test conditions or to compensate for heat-generating DUT. Variable-speed blower systems allow for both situations, although some chambers operate with a constant high flow rate through them regardless of the situation. The dispersion of air flow is important for maintaining homogeneous conditions in the test space. Air

11: ACCELERATED E NVIRONMENTAL C HAMBERS

AND

baffles, deflectors, and diffusion plates are employed in chamber designs to facilitate uniform air flow through the working space. It is important to recognize that even perfectly distributed air flow does not ensure perfect uniformity. When the incoming air has different temperature and humidity levels than the set point, a necessary condition during transitions or when the DUT generate heat, the areas of the working spaces at air intake and exhaust locations must also be different. This has implications for measurement, control, and calibration of temperature and humidity chambers. These points will be discussed throughout this chapter.

11.2.2 Temperature Control Chamber temperature control can be specified for the air or a specific DUT. Controlling temperature of the DUT seems like an easy solution to many testing problems. The opposite is true. Accurate control requires a responsive system, and the mass of many items tested can add a significant amount of latency. This problem often creates significant instability during constant-temperature conditions. Perhaps a better way to deal with the problem is to record the temperature of the DUT but allow the chamber software to control chamber air temperature; then the user adjusts the set point to achieve the desired DUT temperature. Modern chambers can do this automatically, a point discussed later. In the case of PV modules, the safety and performance series of standards published by TC82, IEC 61730 and 61215 respectively, specify the module temperature in the various temperature and humidity tests [1], but in practice many laboratories control the air temperature of their chambers and monitor the module temperature. When the air temperature is specified, it can be measured in the chamber working space, at the point of entry into the working space, or in the chamber exhaust. Regardless of where the thermometer is located, the control system is typically capable of accommodating a user-input temperature offset to compensate for non-uniform conditions and inherent biases between the control sensor and the working space. Calibration will be discussed later, but first is a description of the techniques for controlling temperature and humidity.

11.2.3 Heating and Cooling Heating is achieved by electrical resistance heaters placed in the airstream and/or in water immersion

T ESTING

OF

PV M ODULES

259

systems as part of humidity generation. Heat transfer is usually achieved by forced air convection, but heat can also be radiated from surfaces in the working space. The latter technique is common in chambers designed for the salt mist test, discussed later, but it can also be employed to prevent water from condensing on relatively cool interior surfaces during high temperature and humidity conditions, when insulation is insufficient. Viewing windows are sometimes heated for this reason. The cooling technology is more complex. At its simplest, cooling can be achieved by blowing laboratory ambient air through the chamber without any refrigeration. Such chambers are often referred to as ovens because they operate at a minimum temperature of room ambient plus some offset, say 10e20
C. The IEC safety qualification standard for PV modules, 61730, includes exposures at 105
C “dry” conditions, which can be performed in a controlled oven. If such a chamber includes humidity control, it can be used for the ubiquitous 85
C/85% relative humidity (RH) damp heat test described in the IEC PV module qualification standards and the 60068-267 standard. Of course, cooling below room ambient temperatures is required by many tests for PV modules and other devices. Several required PV module tests include temperatures as low as 40
C. This is an interesting temperature because it is a common line of demarcation between refrigeration systems using a single compressor and “cascade” systems that use two. A brief explanation of the refrigeration cycle is included here to describe the differences between the two. The refrigeration cycle begins with a gaseous refrigerant being compressed into a liquid state by a mechanical compressor into a part of the system called the condenser. Compressing to a higher pressure heats the gas/liquid mixture, and that heat is taken away by a heat exchanger. A valve then allows the gas to expand into an evaporator, where thermodynamic processes immediately cool it to its boiling point. Another heat exchanger draws heat from the space being cooled. The evaporated refrigerant, at low pressure, is drawn into the compressor to repeat the cycle indefinitely. This describes a single compressor system using a refrigerant with a boiling point of approximately 46
C at normal atmospheric pressure. It is this boiling point that impacts the minimum temperature achieved. Although a single compressor could in theory be combined with a refrigerant with a lower boiling point, the pressures

260

D URABILITY AND R ELIABILITY

OF

P OLYMERS

required to condense the gaseous refrigerant would be too high for the piping and valves. When lower temperatures are required, a twostage, or cascade, system is used. Conceptually, a cascade refrigeration system is comprised of two independent single-stage systems joined together by a heat exchanger. In this exchanger, the evaporated refrigerant of the first system is used to cool the compressed refrigerant of the second system rather than directly cooling the air in the working space. The second system uses a refrigerant with a boiling point below 80
C and is capable of achieving working space temperatures as low as approximately 70
C, depending on the DUT and chamber design. The disadvantages of a cascade system are higher complexity and lower energy efficiency due to the extra heat exchange process. Of course, these disadvantages result in higher typical costs. However, most manufacturers, but not all, recommend a cascade system for PV module testing despite the fact that the minimum test temperature of 40
C matches the published minimums of many singlestage systems. PV modules are artificially powered for some tests to simulate their power production and subsequent heat dissipation in sunlight. Because numerous modules are often tested in a single

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

chamber, this puts a significant thermal load on the chamber. A single-stage system may not achieve the minimum module temperature quickly enough or even at all [2]. The cascade systems will achieve faster transitions, a factor that can be utilized to decrease test times. This point will be discussed later. Even lower temperatures are possible with a threestage system, and liquid nitrogen or carbon dioxide can be injected directly into the working space for even lower temperatures or to provide a very rapid cooling of the DUT. However, these technologies are largely irrelevant for testing PV modules, components, or materials, so they will not be discussed here.

11.2.4 Humidity Control Fig. 11.1 shows the relationship between absolute humidity, temperature, and relative humidity (RH). This is a critical point to consider for the discussion of humidity control in environmental test chambers. The mass of water that can be sustained in vapor form in air increases with temperature. For example, at 85
C, RH of 85% means that air, at normal atmospheric pressure, contains 298 g of water vapor per cubic meter (g/m3). This is its absolute humidity. At the same RH, air at 25
C contains only 20 g/m3 of water vapor, or approximately one order of

Figure 11.1 Dew point chart showing the relationship between temperature and relative humidity for several dew points and their respective absolute humidity levels.

11: ACCELERATED E NVIRONMENTAL C HAMBERS

AND

magnitude less. This difference is important to remember when considering the methods available to add humidity within a temperature and humidity chamber, which are discussed below. Another way of representing absolute humidity is the dew point, a useful concept for understanding the relationships between temperature and RH. The dew point represents the temperature at which water vapor in the air will begin to condense if the air were cooled. In other words, the dew point represents the temperature required for air to be saturated, or at 100% RH, given the amount of water vapor currently present in the air. In the examples above, 85
C/85% RH has a dew point of 80.9
C, while 25
C and 85% RH air has a dew point of only 22.3
C. Both dew points are close to the given temperatures because the RH is approaching 100%. The dew point temperature can never be higher than the current temperature of the (A)

(C)

T ESTING

OF

PV M ODULES

261

air, although the two will be equal when the RH equals 100%.

11.3 Methods of Adding Humidity There are four methods for generating humidity inside a chamber. Each has an effect on heat and temperature and can be classified according to whether the process adds or removes heat from the system. The choice of which one to employ largely depends on the desired temperature and RH range for testing.

11.3.1 Steam Generation (Boiler) A boiler is an enclosed water vessel with an immersion heater (Fig. 11.2A). As the water temperature approaches boiling, heated water vapor enters the airstream through a tube. The humidity generated (B)

(D)

Figure 11.2 Examples of humidification devices used in environmental chambers. (A) Boiler type humidifier used in an environmental chamber. (B) Heated water bath humidifier used in a condensation chamber (heating element is beneath the bath). (C) Atomizing spray humidification nozzle. (D) Nebulizer humidification system.

262

D URABILITY AND R ELIABILITY

OF

P OLYMERS

is truly a vapor and requires no additional heat. The main benefit of this type of humidity generation is the ability to inject large amounts of vapor into the air. Because heat has been added during the process, boilers can readily achieve high RH at high temperatures (higher absolute humidity) than other systems, and they take up relatively little space [3]. The addition of heat can be a disadvantage when high RH at low temperatures is required, because the heat must be removed. Boilers can also cause RH oscillation as the immersion heater is cycled on and off [4]. Another disadvantage is that boilers can quickly collect mineral deposits from the evaporated water unless it has been significantly demineralized.

11.3.2 Heated Water Bath Like a boiler, the heated water bath adds heat as well as humidity to the air, but without boiling the water (Fig. 11.2B). Unlike a boiler, the bath is open to the airstream and has a relatively large surface area to facilitate evaporation as air flows over the water. The heater is controlled according to the needs of the system. The responsiveness of the water bath depends on the water’s surface area to volume ratio. Higher ratios result in faster responsiveness and more humidity released into the air, but lower ratios offer more stability [4]. The main advantage of a heated water bath is that it can balance stability and responsiveness at moderate absolute humidity levels, but it may not provide enough humidity to achieve high absolute humidity. It can also take up more space than boilers or other systems due to the need for a large surface area. Another disadvantage is that low absolute humidity levels can be difficult to achieve because the water bath is constantly releasing water vapor into the air stream, even if it is not desired. Heated water baths are the most forgiving type of humidity generator when it comes to harm caused by minerals in the water [3]. No minerals are introduced into the air stream, and buildup on heaters and other surfaces tends to be very gradual.

11.3.3 Atomizing Spray Atomizing spray systems combine compressed air with water at an atomizing nozzle to create very fine droplets (Fig. 11.2C). Usually, the atomizer is located downstream from the chamber heater, and the heat evaporates the droplets as they pass by in the moving air. This technique removes heat through evaporation, which can be advantageous or not depending on

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

the programmed settings. Atomizing spray humidifiers are effective at adding a significant amount of moisture into the air, and they take up very little space. The cooling effect from the droplet evaporation must be compensated for in the heating system, but this means simply increasing the power of the heater. Besides the need for extra heating capacity, the main disadvantage of this method of humidity generation is the possibility of clogging of the spray nozzle with minerals in the water or releasing small particles into the airstream. Supplying compressed air could be another disadvantage in chambers that do not require it for refrigeration.

11.3.4 Ultrasonic Nebulizers Like atomizing spray nozzles, ultrasonic nebulizers remove heat from the system through evaporation (Fig. 11.2D). They are very commonly used for home and healthcare applications where humidity is added to dry room environments for comfort and health reasons. Although normally damaged or destroyed by high temperatures, more robust designs are available for commercial use in environmental chambers. Ultrasonic nebulizers use a piezoelectric material in a circuit that vibrates at a very high frequency, ejecting extremely fine droplets of water from the surface. The nebulizers are kept in a shallow bath of unheated water, and the droplets produced are drawn into the air stream and quickly evaporated. The cooling effect is beneficial when attempting to achieve high humidity at relatively low temperatures, but their use in chambers requires additional heating capacity to compensate for this effect. Another benefit is that this technology does not require the use of compressed air, making it energy efficient and quiet. The main disadvantage of ultrasonic nebulizers is that they have to be spread over a relatively large surface area to generate significant amounts of humidity, much like heated water baths. This drawback tends to limit their use to small chamber sizes.

11.4 Methods of Removing Humidity Removing humidity can be just as important for some test conditions as adding it. The two most common methods for this are refrigeration, used by almost all chambers, and the use of desiccants in two types of systems [5]. It is important to recall the previous discussion concerning absolute humidity

11: ACCELERATED E NVIRONMENTAL C HAMBERS

AND

and dew point. Without the ability to remove humidity from the air, the chamber RH is affected by laboratory conditions. Reducing RH, in this case, can only be achieved by heating the air, but what if you want to achieve low RH and low temperature? Chambers must include one of the humidity-reducing technologies to make this possible. The most common technique for humidity removal is refrigeration because it also cools the air, achieving two goals simultaneously. When air is cooled, it is less able to sustain water vapor and condensation occurs; water condensed in this manner in a chamber is drained away. Typical chambers are capable of achieving a dew point of approximately 7
C. This value is derived from the efficiency of heat transfer in an exchanger and the fact that cooling coils typically must be kept above the freezing point of water to avoid buildup of ice. Based on the minimum dew point, chamber manufacturers publish temperature and RH specifications in a graphical format called a climatogram, showing temperature on one axis and RH on the other, with one or more shaded areas depicting the chamber’s capabilities. See Fig.11.3 for an example. Desiccants are substances that absorb moisture from the air; silica gel bags commonly used in product packaging are a common example. They are used in two methods for humidity reduction in chambers. In the first, called dry air purge,

T ESTING

OF

PV M ODULES

263

compressed air is passed through a container of a desiccant material on its way into the chamber. Two containers are used, so that one is being vented to room conditions to regenerate the desiccant material, while the other is actively removing moisture from the compressed air. The drying effect of the compressed air decreases the humidity, but additionally this technique is often combined with refrigeration in a specific way. Cooling coils are normally kept above the freezing point of water to avoid ice formation on them, which decreases their heat exchanging efficiency and puts stress on the compressor. However, dry air purge creates a slight positive pressure in the system so that ice sublimates from the coils in the dry environment, allowing them to operate a few degrees below 0
C. This feature is commonly called a “low RH package” or something similar. Large chambers may also include a recirculating desiccant dryer, where air is continuously circulated through a system with a wheel made of a desiccant material. This wheel is partially exposed to room conditions or heated to release moisture for regeneration. A third type of dehumidification, nitrogen purge, is used by some chambers, but typically not those involved in testing PV modules. Nitrogen gas does not retain water vapor, so it can be introduced into the working space to push moisture-laden air out. The technique is hazardous to people because it drives oxygen out of the working space, so it is rarely used

Figure 11.3 Climatogram showing the achievable temperature and relative humidity settings for a corrosion chamber with and without a refrigeration module.

264

D URABILITY AND R ELIABILITY

OF

P OLYMERS

in a walk-in type chamber. For PV module testing, the extreme low humidity levels possible with nitrogen purge are unnecessary.

11.5 Measuring Relative Humidity in Environmental Chambers Measuring relative humidity is surprisingly difficult. Many techniques and technologies have been developed over the centuries, the oldest known tracing back to the Western Han dynasty in China around 200 BC [6]. Early instruments used horse or even human hair in tension in a dial-type gauge. Leonardo da Vinci developed a crude type that was refined in the 1700s. Fig. 11.4 shows an example of a 19th century device still in operation in Salzburg, Austria. Today’s state-of-the-art measurement systems are based on a principle developed by the National Institute of Standards and Technology (NIST). This technique starts with fully saturated air at a known pressure (higher than atmospheric), and temperature. The air is allowed to expand through a valve to normal atmospheric pressure, where the temperature and pressure are again measured and the

Figure 11.4 A 19th century weather station hygrometer in Salzburg, Austria.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

relative humidity of the new environment is known by applying thermodynamic principles. Effectively, the technique involves a very special climatic chamber that generates a precisely known relative humidity, which then is used as a reference for calibration of other devices. RH measurement is a subject worthy of extended discussion, but for the purposes here, a summary of three techniques will be discussed in the context of their application to environmental chambers and PV module and material testing. All three types of sensors, shown in Fig. 11.5, are used for direct measurement and control within the chamber or for calibration. There are certainly other measurement methods, and some of them could be applied to environmental chambers.

11.5.1 Psychrometric Wet Bulb/ Dry Bulb Sensors The earliest fundamental measurement technique for RH, wet bulb/dry bulb sensors, continue to be commonly used today. These sensors rely on two thermometers. One is positioned in the air as any thermometer might be, while the other includes a wick or sock over the end which is kept constantly wetted. Evaporation from the end of the wick depresses the temperature of that sensor, and this difference is used to determine the relative humidity from thermodynamic principles. At 100% RH, there is no evaporation from the wick, so the two temperatures are equal. In practice, these formulae are input into computer software to perform the calculations automatically, or the software may simply include well-developed data tables to relate the two measured temperatures to an RH value. The latter technique is most common because the use of thermodynamic equations must include either a measurement of atmospheric pressure or an assumed value. The data tables assume the atmospheric pressure, and these are accurate in most instances. As with any measurement technique, this technique has strengths and weaknesses. The benefits of psychrometric RH measurement systems are that they can be very accurate, since they rely only on accurate measurement of temperature. These sensors tend to be very responsive and robust, although they do require some maintenance. Wicks need to be replaced periodically, and the volume and purity of water keeping the wet bulb wet must be maintained. Wet bulbs drying out and contaminated wicks are the

11: ACCELERATED E NVIRONMENTAL C HAMBERS

AND

T ESTING

OF

PV M ODULES

265

Figure 11.5 The three main types of RH measurement devices used in environmental test chambers.

major sources of measurement error in these systems. In either case, evaporation is inhibited or eliminated, and the RH measurement is biased in the direction of saturation. If a dry bulb becomes wet, the same bias occurs, although this is typically only a temporary condition if it occurs. Another limitation is that the placement of such sensors is limited to locations where water tubing can be easily routed.

11.5.2 Electronic Sensors Certain materials change electrical impedance, either resistance or capacitance, as a function of humidity. These properties can be measured and correlated to the relative humidity of the environment; therefore these are secondary, or indirect, measurement techniques. If the chamber requires RH measurements of air below 0
C, these can be the only practical solution. Additionally, certain situations may make the maintenance of a wet bulb wick difficult or impractical, and electronic sensors are easier to use. However, electronic sensors are susceptible to contamination, to a greater extent than wet

bulb wicks. Although it is possible to protect them with the use of semipermeable materials, this technique slows down the response time of the sensor and may introduce hysteresis into the system. Another problem is that many electronic sensors have difficulty operating at a high RH, and failures of this sort are not easily detectable. Nevertheless, electronic sensors offer sufficient accuracy for most applications and are commonly used.

11.5.3 Chilled Mirror Dewpoint Hygrometer Another fundamental RH measurement technique involves detecting the dew point in the environment. In theory, the chilled mirror technique can be used to detect condensation of any gas, but for the purposes of temperature and humidity chambers, it is used to detect condensation of water and thus the dew point of the air. A chilled mirror hygrometer is composed of a thermal conducting metallic mirror surface, a thermoelectric cooler, and a platinum resistance thermometer embedded in the mirror. Also included

266

D URABILITY AND R ELIABILITY

OF

P OLYMERS

is an LED, a photodetector, and a separate platinum resistance thermometer. A beam of light from the LED is reflected off the mirror surface and received by the detector. The temperature of the mirror surface is reduced until condensation commences. As dew is formed on the mirror, some of the light is scattered, and the detector output decreases. The system then iteratively adjusts the temperature of the mirror surface until it detects the precise temperature at which equilibrium is achieved, which is when the amount of condensation on the surface is neither increasing nor decreasing. This is the dew point. A separate temperature measurement of the air is all that is needed to accurately calculate the relative humidity. This is perhaps the most accurate technique available in a device small enough to be placed in most chambers, although small psychrometric devices do exist. One drawback is that the mirror surface must remain very clean for the technique to work, making it less practical for use as a chamber’s control sensor. Another is that the iterative determination of the dew point introduces some latency into the measurement, especially in a rapidly changing environment such as a chamber during a transition. However, chilled mirror hygrometers are used for calibration of electronic sensors and to measure the RH for the purposes of determining chamber uniformity.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

system often. Transitions are the next subject discussed.

11.6.1 Transitions Between Test Conditions: Step Versus Linear

11.6 Temperature and RH Control: Putting It all Together

Transitioning between conditions is one of the most important aspects of environmental testing. The IEC qualification standards for PV modules include two different tests featuring temperature and/ or humidity transitions. Yet transitions are one of the most variable parts of any test, and they pose challenges for laboratories and chamber manufacturers alike. There are two types of transitions in an environmental chamber, see Fig.11.6. First is a step transition. A step transition says nothing about how long it should take to attain the condition being transitioned to. It says, go from condition “A” to condition “B.” Many standards in temperature and humidity testing, weathering, and corrosion include these. It seems likely that in some cases the inclusion of a step transition is unintentional. Many readers of these methods interpret a step transition to mean, “Get from A to B as quickly as the chamber can do so, but as long as it gets there by the end of the step, that is acceptable.” Other readers believe the transition should be instantaneous, as in a thermal shock chamber that includes two working spaces, one hot, one cold, and a lifting system to move specimens quickly between them. Some standard test methods provide guidance by giving a time limit on the

Temperature and humidity control systems are founded on principles of proportional, integral, and (sometimes) derivative (PID) control algorithms, discussion of which is beyond the scope of this chapter. Chamber manufacturers expend significant resources perfecting their control software and consider this proprietary information. However, common techniques present themselves and are openly discussed by some manufacturers. Chambers that are designed to operate over a wide temperature and RH range may begin by chilling and drying air rather than rely on ambient conditions in the laboratory. Heat and humidity are then added back into the system for precise control since these processes are easier to administer in small doses. High airflow is required when attempting to accomplish precisely-controlled transitions because latency is minimized and more opportunities exist to condition the air appropriately when it recirculates through the

Figure 11.6 Linear and step transitions between conditions in an environmental test.

11: ACCELERATED E NVIRONMENTAL C HAMBERS

AND

transition. Occasionally, this ambiguity is intentional, particularly when there are factors to performing a test external to chamber capabilities that limit the ability of standards writers to provide specific guidance. Regardless of the reason, step transitions often lead to disagreements between people who test, people who request testing, and manufacturers of the chambers who perform tests. The second type of transition is a linear one where a fixed rate of change over time is directly specified or implied. Effectively, a linear transition serves to slow down transitions compared to a typical step transition. This is because most chamber software is programmed to see a step transition as a “get there as fast as possible” goal and works the heating and cooling hardware near its limits to achieve condition “B” quickly. There are several reasons to slow down a transition. One is to reduce variability between tests and make it clear to all parties how the transition is to occur. By controlling the rate of the transition, differences between chamber heating and cooling capacity are mitigated, as are differences in heat produced by various DUT. The latter difference can be quite significant when testing powered PV modules. Another reason to slow the transition is to avoid unrealistic stresses caused by rapid heating and cooling. A PV module in service can be subjected to wide variations in temperature, but these occur relatively slowly in the natural environment. Designers of the PV module qualification test sequence intended that PV modules be subjected to a thermal cycling test, not a thermal shock test. Yet a third reason to slow transitions is to regulate the time in specific temperature or RH ranges, a consideration especially important in corrosion testing or when the glass transition temperature of a polymer being tested is a concern. In the case of step transitions, chamber manufacturers have a competing challenge to “getting there as fast as possible.” Severe temperature overshoots1 are usually not desirable, so control algorithms are designed to slow down transition rates as the set point is approached. Chamber manufacturers typically make great efforts to minimize the temperature or humidity overshoots while transitioning. However,

1

Overshoot includes situations when the condition temporarily goes below the set point during a transition to a lower temperature or RH, not just above the set point during transitions upward.

T ESTING

OF

PV M ODULES

267

this can be a problem when a test method uses a step transition and also specifies a DUT temperature. A massive object takes longer to achieve condition “B” in a transition than the chamber air temperature, but air temperature is what is traditionally measured by a chamber. In response to this issue, many chambers have the ability to place a temperature sensor on or inside a DUT and control the chamber to that sensor. Some modern chambers can be programmed to allow the air temperature to go beyond the set point of the next step to speed up the heating or cooling of the DUT. When the DUT temperature approaches the set point, the air temperature is adjusted to gradually equilibrate. This “controlled overshoot” feature is especially useful in tests that require a minimum or specific “soak” or “dwell” time for the DUT after one transition before moving on to the next one. This is the case in the PV qualification standards.

11.6.2 Characterizing Environmental Chambers: Uniformity, Fluctuation (Stability), and Heating and Cooling Rates Environmental chamber specification sheets include performance parameters such as minimum and maximum temperature and relative humidity, heating and cooling rates, and sometimes uniformity of conditions in the working space. Much of this information is given according to standard characterization techniques described in IEC TC104 standards, and these will be briefly described here. IEC 60068-3-5 is a standard on characterizing temperature chambers, while its companion 60068-36 characterizes humidity chambers. These standards include a test sequence, data collection regime, and calculation criteria to characterize the uniformity of the chamber, referred to as gradient or variation in space, fluctuation of conditions over time, and heating and cooling rates. In the procedure, temperature sensors are placed throughout the chamber in defined areas, at least nine of them, and data is recorded during a test sequence that includes transitions to the high and low temperatures of the chamber’s operating range and operation in stable conditions. This data is used for a variety of characterization calculations and other purposes as well. These are worth discussing. According to the standard, two terms describe uniformity. Gradient represents the difference

268

D URABILITY AND R ELIABILITY

OF

P OLYMERS

between the maximum and minimum temperatures of all the positions measured at constant-temperature conditions. Variation in space refers to the largest difference between the center and any of the other measured positions. Some chamber manufacturers report the uniformity as plus/minus the standard deviation from the mean of all positions measured. All three methods can be relevant depending on exactly what is being tested inside a chamber. Of course, once specimens are in the chamber, the air flow changes, possibly impacting the uniformity numbers generated under empty chamber conditions. Fluctuation, referred to in some weathering and corrosion standards as operational fluctuation, is calculated as plus/minus two standard deviations from the mean. It is a measure of the stability of conditions while operating at a steady state. To generate this data, the chamber is allowed to stabilize at a constant-temperature setting. Then, data is recorded at regular intervals, usually once per minute or more, for 30 min or longer. This data is averaged and the sample standard deviation calculated. It is important to note that the standard technique performs this test with an empty chamber. When using a DUT temperature control, massive specimens can introduce latency and increase fluctuation. Heating and cooling rates are very important for many applications. The standard method for calculating heating and cooling rates is to perform step transitions from minimum to maximum temperature, in both directions, and measure the time it takes to achieve each one. However, because control algorithms will slow heating and cooling processes as the set points are approached, the standard says to eliminate the time it takes to move through the first and last 10% of the operating range. This provides a rate in terms of Kelvin or degrees Celsius per minute. Some manufacturers also publish transition times between the two temperature extremes, with or without a specimen load. Some publish data showing the transition rates with various specimen loadings by mass. Faster rates require more heating and cooling power, which increases costs. The heat generated by the DUT is significant for PV modules tested while supplied with current. For this reason, testing PV modules typically requires a chamber with a faster cooling rate than specified by the climatic test. However, the rate of change in the PV standards is low relative to the capabilities of many chambers. Linear control is required in some cases to slow down

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

the transition to comply with test requirements, which will be described later. Another use for this data is calibration and inputting offsets into control software. This can be necessary due to the location of control sensors. For practical reasons, control sensors are often outside the working space. Even if located within the working space, typically they are not in the center. The data collected in the 60068-3-5 standard sequence can be used to calibrate according to the needs of the test application. In some cases, operators may choose to take the average of all positions in the chamber as the reference temperature against which an offset is input into the controller. In other cases, the center value is used. DUT or product temperature control is another way to address this issue because the control sensor is directly affixed to the specimen or a representative specimen under test, albeit with some associated problems discussed previously. The PV module standards require measuring the temperature of the module, or a single “representative” module in a chamber filled with multiple specimens. The IEC committee devotes an entire standard to characterizing humidity chambers, but the 60068-3-6 standard is nearly identical to its temperature counterpart. However, there is one very important difference. Difficulties with RH measurement were discussed earlier. In recognition of this fact, the IEC procedure includes a single RH measurement in the center of the chamber2. With the RH and temperature measured in this location, the absolute humidity is known. The standard goes on to make the assumption that the absolute humidity is perfectly uniform throughout the chamber. The other temperature data points can be applied to this absolute humidity to calculate the RH in each of those locations. As an example, suppose one is performing this measurement in a chamber running the damp heat test, 85
C at 85% RH, and the center measures exactly these conditions. Elsewhere in the chamber, one

2

Although the standard says the measurement is to be taken at the center of the working space, in some cases it is better to place the RH sensor elsewhere. A good example is a corrosion chamber with the spray nozzle in the center, a very common arrangement. In this case, the center of the chamber is not part of the working space. Moving it somewhere in the usable test space is sound engineering judgment. Words in a standard should not override good practice.

11: ACCELERATED E NVIRONMENTAL C HAMBERS

AND

thermometer records 83
C while at the opposite extreme it is 87
C. Applying the absolute humidity of 298 g/m3 to these temperatures yields RH values of 92% and 79% respectively3. Water vapor diffuses quickly through a mass of air, so the assumption is reasonable. This is a very practical solution to the problem. RH sensors with high enough accuracy are expensive and rather large. Mounting nine of them in a chamber would be exceedingly costly and difficult.

11.7 Salt Mist and Corrosion Chambers The standard qualification sequence does not require any salt mist corrosion test. However, references to it are made for PV modules installed in corrosive environments, such as near the ocean or roads that are salted for snow and ice removal. Salt mist, or fog, testing has existed for more than 100 years. This is sometimes considered a special case of environmental testing, and today corrosion and environmental testing are merging in significant ways. The basic salt fog test, described in ASTM B117, ISO 9227, and IEC 60068-2-11, includes a 5% sodium chloride solution pumped to an atomizing spray nozzle and combined with compressed air to create a fine mist. This process is essentially the same as atomizing spray humidity generators, except with a corrosive solution instead of purified water. In fact, there are test standards for paints that use an atomizing spray nozzle, called a fog nozzle, to generate saturated humidity to assess moisture resistance. Most salt fog testing is done at a constant chamber air temperature of 35
C. In some cases, salt mist tests are conducted on electrotechnical devices as an

3

There is a great deal to be said about stated tolerances in test standards, but this example shows why many of them have no practical basis. The tolerance for damp heat typically has a plus/minus of 5% RH. According to the IEC 60068-2 series of standards, tolerances incorporate all factors of error, including measurement error, fluctuation, and uniformity. If one properly accounts for measurement uncertainty, which is 2% to 3% in ideal conditions [10], then all of the other factors must be limited to about half of the stated tolerance, or about 2.5% RH. This is all but impossible in a real testing situation. This problem unnecessarily consumes a lot of time in the testing industry.

T ESTING

OF

PV M ODULES

269

accelerated moisture ingress test. When in a typical humidity chamber, any water that permeates or penetrates an enclosure may not immediately cause damage to electrical components because of its purity. Adding salt to the environment ensures rapid damage to components should the mist make its way inside. Over time, the salt fog test has been combined with climatic testing, starting with simple two-step tests where periods of fog were alternated with forced-air drying by heated ambient air [7]. Later, saturated humidity steps were added into these test sequences. RH was generated using all of the techniques discussed earlier with the exception of ultrasonic nebulizers. RH control was rudimentary, however, and conditions were either wet or dry without any control in-between. These tests, commonly called cyclic corrosion tests, proved inadequate and were prone to a variety of problems, including poor reproducibility. As a consequence, many of today’s automotive corrosion test methods include much of the language and methods of environmental tests, including RH control and linear ramping between conditions [8]. Environmental testing with salt mist includes several cyclic test methods in IEC 60068-2-52, which are referenced in the PV standard IEC 61701. These methods mirror early attempts to improve the tests by moving specimens from a salt spray chamber into a temperature and humidity chamber. These tests are labor-intensive and prone to variability due to the manual handling of specimens. However, in the past there were no other options because controlling the environment in the same chamber as salt spray had not been reliably achieved. Today, this has changed because of improved designs, but several challenges must be overcome. Controlling relative humidity in a salt spray environment is extremely challenging because all RH measurement technologies are susceptible to temporary or permanent damage when exposed to salt spray. Techniques to mitigate this problem include retractable sensorsdwhich are inserted during climate control and removed from the working space during periods of fogdand sensor placement in the recirculating air stream outside the working space. Temperature and humidity chambers do not normally have significant volumes of liquid water in the working space that must be evaporated before decreases in RH can be attained. In a salt spray environment, this is the primary part of the test. Additionally, salt fog chambers tend to have salt

270

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Figure 11.7 Painted metal specimens loaded into a horizontally designed salt spray chamber.

deposits accumulating in the working space, which serve as large desiccant blocks which remove humidity from the air. Some chambers include hardware to wash surfaces of deposited salt [9]. Another challenge salt fog chambers face is that they must be designed horizontally so that specimens do not drip onto one another (see Fig. 11.7). Corrosion products from one specimen would contaminate others if this were to occur. This disadvantages the common cuboid shapes that promote uniformity because specimens can only be placed in a single horizontal plane within the large space. Although seemingly a minor issue, the spray hardware obstructs air flow, creating an extra challenge to achieving uniformity. More significantly, corrosion chambers must rely, at least partially, on radiant heat because flowing air through the chamber dramatically disrupts the salt fog dispersion. Salt fog flowing through recirculation systems will damage components such as heaters, blowers, and sensors. These unique challenges for corrosion testers add up to somewhat compromised uniformity, fluctuation, and ramp rates compared to comparably-sized climatic chambers. The benefits of automatic operation and reduced specimen handling should be considered in balance with these trade-offs.

11.8 Weathering Chambers A definition of weathering is provided in this chapter’s introduction. The key factor is that weathering degradation starts with photons of light disrupting chemical structures, followed by heat and moisture processes further degrading the material. Weathering testing is not new to the PV industry.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Polymeric materials used for backsheets, encapsulants, and edge seals have been supplied by companies with experience in weathering testing from other industries they sell into. These suppliers, along with testing organizations and chamber manufacturers, spearheaded the IEC standards for testing components. Two technologies have been recognized for standardized testing, and they are the focus of this section. Fluorescent UV and xenon arc weathering chambers will be treated as special cases of climatic chambers. What sets weathering chambers apart is the existence of a light source(s) as the central component of the chamber, around which all other climatic factors have to be designed. One consequence of this fact is that the weathering chamber working spaces are mostly planar surfaces rather than the three-dimensional volumes of temperature and humidity chambers.

11.8.1 Fluorescent UV Weathering Chambers In weathering, fluorescent UV chambers are probably the most common type used. The technology was introduced in 1970 and standardized later in the decade. It is most commonly used for testing paints and plastic resins, but many other applications exist. The fluorescent lamps used are the same as standard fluorescent lighting lamps, except that their phosphors fluoresce a broad UV spectrum rather than in the visible region. Two main types exist and have been described in ASTM G154 and ISO 4892-3: UVA-340 and UVB-313. Each emits energy in both the UVA and UVB regions, but their peaks and cut-on wavelengths are different. Fig. 11.8 shows spectral irradiance distributions for these two types along with the ASTM G177 peak natural sunlight standard4. This sunlight spectrum has higher irradiance than what the PV industry considers “one sun.” To illustrate this, at 340 nm, a common control wavelength for weathering devices, the spectral irradiance is 0.73 W/(m2 nm), while the standard preferred by the PV industry has a value of 0.50 W/(m2 nm) at 340 nm. The charted spectral curves clearly show that the UVA-340 fluorescent lamp has a very good 4

The G177 standard includes a published spectral irradiance table only in the UV, but generating a full spectrum chart is a simple matter of inputting the G177 parameters into the SMARTS2 model.

11: ACCELERATED E NVIRONMENTAL C HAMBERS

AND

Figure 11.8 Spectral irradiance curves of standard fluorescent UV lamps versus peak sunlight.

match to peak sunlight in the entire UVB range and most of the UVA range. For wavelengths greater than approximately 360 nm, the irradiance drops off and is deficient compared to sunlight. UVB-313 lamps emit shorter wavelengths than natural sunlight. In addition to low heat radiation, a key benefit of fluorescent UV lamps is their spectral stability. As the lamps age, the irradiance output decreases but the shape of the spectral curve does not. Variable power supplies and feedback irradiance control systems compensate for reduced irradiance, and a lamp life at peak summer sunlight levels can be thousands of hours. The most common chambers can operate at “three sun” irradiance levels at 340 nm, although the lamp life is reduced. The fluorescent UV lamp technology is the most economical of any method used for weathering. Fluorescent UV weathering chambers are unique among environmental chambers. The working space of the chamber is not in the interior of the chamber, but rather on the two outer walls that run parallel to the two banks of lamps. These walls are formed by inward-facing holders for flat specimens and not insulated. Temperature is controlled by a thermometer affixed to a black metal panel, similar to DUT control in climatic chambers. The fluorescent lamps radiate little heat, and heated forced air convection is needed to achieve most set points. The chamber includes a heated water bath with temperature control. During the condensation function, the lamps are turned off and the backs of specimens are cooled by room air. As the temperature increases in the chamber interior, specimens radiate heat to the outside and remain below the dew point of the humid air inside, creating continuous condensation on the specimen

T ESTING

OF

PV M ODULES

271

surface. The technology is very simple, but heated condensation as a stressor is unique in the field of climatic testing. As an independent test, “hot condensation” is one of the first tests performed on new paint formulations or for quality control. Blisters form quickly if the paint lacks resistance to moisture. Tests in fluorescent UV chambers typically alternate between UV exposure in dry, heated conditions and condensation. Cycles are normally programmed in blocks of 4, 8, or 12 h. Because condensation formation can take as long as an hour to occur due to the slow heating of the water bath, a minimum recommended time for a condensation step is 4 h. Water spray is available in these chambers, providing some mechanical washing of the specimen surface and a mild thermal shock to specimens from the relatively cold temperature of the spray. Thick or thermally insulating materials may not become adequately wetted by condensation because heat cannot sufficiently radiate to the outside room, preventing the front side temperature from decreasing much below the dew point. Water spray can mitigate the problem by wetting and cooling the specimen surface. The general principle for using fluorescent UV lamps to conduct weathering tests is that UV light is responsible for the vast majority of photodegradation of polymers, especially those considered durable enough for outdoor use. The energy produced by this system is concentrated in the spectral area of interest for most testing. This efficiency means that there is no need to remove excessive heat produced by the lamps, simplifying the overall chamber design. UV conditioning tests for PV modules represent a special case of fluorescent UV chambers. The qualification tests, known as MQT 10 in IEC 61215-2, describe exposure dosages in the UVB and UVA ranges but not what light source is to be used to achieve them. Although xenon arc and metal halide chambers can be used, fluorescent UV lamps may be the most practical. ASTM E3006 describes a method to achieve the requirements of the UV preconditioning tests in the IEC qualification standards using standard fluorescent UV weathering lamps and control technology. Original exposure dosages required a mix of UVB-313 and UVA-340 lamps or specially designed lamps for the test. Updated requirements have changed the dosage so that the UVB-313 lamp is no longer needed. Fluorescent UV weathering chambers are not large enough for module exposures, but specially constructed chambers have existed for many years. Several manufacturers

272

D URABILITY AND R ELIABILITY

OF

P OLYMERS

Standard Fluorescen nt UV weathering chamber

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

UV condioning test chamber from WeissTechnik

Figure 11.9 Photographs of a standard fluorescent UV chamber (Q-Lab Corporation) and a commerciallyavailable chamber designed to perform UV conditioning tests in the IEC qualification standards for PV modules. Courtesy of WeissTechnik GmbH.

of temperature and humidity chambers have fitted their products with banks of fluorescent UV lamps to perform these tests. See Fig. 11.9 for photos of standard fluorescent UV chamber and one designed for PV module testing.

11.8.2 Xenon Arc Chambers Xenon arc lamps were invented in the 1950s in Germany and began to be used almost immediately for weathering testing. These lamps are made from a quartz tube (long arc) or bulb (short arc) with various pressure levels of pure xenon gas inside them. Long arc lamps are the type used for weathering because of their relatively long life under continuous use. The arc emits a continuous spectrum that approximates sunlight in the UV, visible, and IR spectral regions. Today the technology is the most commonlyused light source for simulation of the full spectrum of sunlight. Xenon arc lamps produce shorter wavelengths than terrestrial sunlight, into the UVC range. Therefore optical filtration is used to accurately simulate the UV spectrum. In addition, xenon arc lamps have very high irradiance peaks in the near infrared region. This energy results in extra radiant heat, beyond what a perfect sunlight simulation would produce, and this must be removed by the air handling system unless the temperature set point is very high. Some optical filter systems use infrared-absorbing coatings or reflecting mirrors to reduce the radiant heat load in the working space. In any case, this heat must be removed by the

system. Xenon arc lamps require cooling to prolong their service life, and this can be achieved by either water or air circulation. The choice of which type to use is dictated by the power density of the lamp. Higher wattages in a smaller package require water cooling because air cannot remove heat quickly enough. When air cooling is sufficient, it is the preferred choice because of the added complexity of water cooling systems. Xenon arc lamps have a limited useful life because aging of the quartz tube and degradation of electrode materials cause the UV output efficiency to decrease. At around 1500 h, the UV output has decreased enough that replacement is necessary. At low irradiance levels the useful life can be extended somewhat. Perhaps the most critical variable in xenon arc testing is the choice of optical filters to use. These remove short wavelength energy from the working space. They can be characterized by a “cut-on” wavelength which is the shortest wavelength with irradiance above a nominal threshold value. Weathering standards have defined three families of filtersdDaylight, Window, and Extended UVdand within each family is a multitude of chamber specific and proprietary types. Daylight filters are designed to simulate direct sunlight, which has a cut-on wavelength of around 295 nm. Window filters simulate sunlight through a vehicle or building window, which filters most or all of the UVB. Extended UV filters transmit some shorter wavelengths than experienced from natural sunlight. Seemingly small differences between optical filters sometimes result in significant

11: ACCELERATED E NVIRONMENTAL C HAMBERS

AND

differences in material degradation. This creates reproducibility problems between laboratories, which is common. For this reason, some newer test methods have been written with a narrowly-defined UV spectral irradiance. The older definitions of Daylight and Window filters attempted to standardize around the many types offered by xenon arc chamber manufactures, but the newer definitions attempt to focus the range. The IEC weathering standard for PV materials uses an optical filter definition developed for ASTM D7869, a standard for automotive paints and exterior components. The IEC standard also uses the same maximum irradiance, 0.80 W/(m2$nm) at 340 nm, which is slightly higher than peak sunlight (see Fig. 11.10). Acceleration comes from the fact that this peak irradiance is run throughout the entire cycle, day in and day out. Testing near the maximum operating temperature also provides acceleration. Climatic control in xenon arc weathering chambers is more challenging than other types of chambers because of the significant radiant heat produced by the lamps. Xenon arc lamps used in weathering chambers range in power from 1500 to 15,000 W, and roughly half of the radiation produced is in the infrared. For this reason, xenon arc chambers control two temperatures in most cases. The first is a black panel temperature or the surface temperature of a thin panel of stainless steel painted black. Two versions of this type of thermometer exist, with the difference being that one of them is affixed to a piece of white plastic to insulate the back side and increase its heat load and temperature. The two are known as

Figure 11.10 Spectra of peak sunlight and the xenon arc spectrum specified in IEC component weathering standards.

T ESTING

OF

PV M ODULES

273

uninsulated and insulated black panel thermometers, with the insulated type also called a black standard thermometer to indicate there are none more black in terms of absorption of thermal radiation. The black panel temperature is a function of the radiant heat load, which is a function of lamp irradiance, the presence of any infrared absorbers or reflectors in the optical filter system, and the chamber air temperature, which is also controlled. The black panel temperature is intended to be an approximation of the hottest specimens in the specimen plane, uninsulated for metallic specimens and insulated for plastics. The two interdependent temperatures are independently controlled by adjusting the ratio of fresh room air to recirculated air and overall rate of air flow that moves through the working space. The overall rate of air flow removes heat from the black panel and controls its temperature, and the ratio of cool room air to conditioned chamber air affects the chamber air temperature. Most xenon arc chambers are equipped with RH control. Humidity is generated with boilers, atomizing spray nozzles, and ultrasonic nebulizers, depending on the chamber size and specific design. Generally, humidity removal is not a concern for xenon arc weathering tests because no standard test methods require low dew points. Typical xenon arc chambers specify air temperatures ranging from a minimum of 38
C to a high as 80
C. The minimum can be achieved by cooling with ambient air and humidity generating techniques that remove heat. Although not common, some chambers have optional refrigeration to run at reduced chamber air temperatures, but not as low as the freezing point of water. Xenon arc weathering chambers are designed in two form factors, rotating rack and flat array (Fig. 11.11). As mentioned earlier, weathering chambers have an essentially planar working space due to the necessity of maintaining a fixed distance between specimens and lamps for irradiance control and uniformity. The flat array form factor does accommodate three-dimensional items, but if the top surface protrudes above the tray significantly, the irradiance will be higher than the chamber setting and the specimen will be hotter, sometimes several degrees hotter. Rotating rack chambers may appear three-dimensional as the specimen rack moves chambers around the central light source(s), but the cylindrical basis of the rack is a planar surface curved onto itself. In short, xenon arc weathering chambers

274

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 11.11 Examples of rotating rack and flat array xenon arc weathering chambers.

11.8.3 Metal Halide Light Sources Metal halide lamps are quartz tubes filled with trace amounts of mercury and metal halide salts. An electrical current vaporizes the mercury and creates spikes of light, mostly in the UV. When the mercury arc is struck, the heat vaporizes the metal halides, and this process broadens the spectrum of light produced. The precise blend of these salts is a key factor in their spectral output. In the field of PV and material testing, there are two uses for this technology. In the first, high power metal halide lamps designed to radiate mostly in the UV can generate high UV

irradiance for materials testing. Second, the salts can be blended so that the lamp produces a broad spectrum similar to sunlight. Fig. 11.12 shows the spectrum of a metal halide lamp used by NIST in its SPHERE weathering chamber, while Fig. 11.13 shows the SPHERE device itself. Although the NIST device is a unique scientific test apparatus, efforts are underway to commercialize a smaller version of it, but for now this is still a technology for experimentation. There are, however, commerciallyavailable metal halide chambers with a similar light source and aspects of climatic chambers. These purport to produce irradiance levels up to 30 times higher than “one sun” levels. However, their match to natural sunlight is not as accurate as the

6.0 5.0 Irradiance (W/m2/nm)

contain a mostly two-dimensional working space, unlike temperature and humidity chambers. Xenon arc chambers include water spray for testing outdoor materials as the only method for wetting specimens5. In the rotating rack form factor, the spray nozzle is turned on continuously during that portion of the cycle, and specimens are sprayed intermittently as they move past the fixed nozzles. Flat array chamber nozzle systems spray all specimens simultaneously, but mimic the rotating rack behavior by pulsing the nozzles on and off during the spray period.

NIST SPHERE Peak Sunlight

4.0 3.0 2.0

5

Some chambers include spray nozzles that spray the back sides of specimens. The intent is to operate the chamber at high RH and cool the specimens by spraying the back side to create condensation on the front side. This has not been proven effective because the spray water is rapidly equilibrated with the chamber air temperature, so the specimen temperature never drops below the dew point.

1.0 0.0 250 300 350 400 450 500 550 600 650 700 750 Wavelength (nm)

Figure 11.12 Spectrum of a NIST SPHERE metal halide lamp.

11: ACCELERATED E NVIRONMENTAL C HAMBERS

AND

Figure 11.13 NIST SPHERE test chamber.

more common weathering chamber light sources. As noted above when discussing optical filters in xenon arc chambers, small differences in the UV spectrum can have significant impact on degradation and test results. Even though most of the energy produced is in the UV, the irradiance levels produce significant heat. These high irradiance chambers usually have liquid-cooled specimen trays to avoid thermal failures of the specimens. Care must be taken to avoid reciprocity assumptions when testing with these chambers, beyond the spectral mismatch issues that exist between metal halide lamps and sunlight. Little evidence exists that the same results are obtained from equivalent UV dosages of identical spectral output, applied at different rates over time. The material behavior to high irradiance versus real-life levels is highly variable according to its chemical structure and other factors. More complex polymer systems are less likely to adhere to linear reciprocity. Nevertheless, quickly determining how polymeric materials degrade from UV light can be useful to experienced material experts. The second type of metal halide chambers uses full-spectrum lamps designed to simulate sunlight. These are commonly used to test for thermal effects

T ESTING

OF

PV M ODULES

275

from solar radiation rather than for weathering. The spectral stability in the UV as lamps age is not well publicized, and irradiance control is often limited to regulating the distance between lamp arrays and the DUT. Large chambers with several lamps mounted in parabolic reflectors can test fullsize PV modules or whole vehicles. IEC 60068-2-5 includes such a test, which is more closely aligned with traditional environmental testing than weathering. In theory, xenon arc lamps can be used for these exposures, and often they are, but the infrared output is excessive and may not thermally stress the materials the same way that sunlight does. It is also very difficult to make large scale xenon arc chambers for something as large as a PV module. Confusion exists about the IEC standard as well as MIL-STD-810G. Makers of electrotechnical devices, including PV modules, have taken these solar radiation thermal stress tests to be weathering tests. TC104 has corrected this misunderstanding by revising 60068-2-5 to specify that weathering is different and is best conducted in a standard weathering chamber. Test “Sa” covers the traditional solar radiation test while the new “Sb” refers to ISO 4892-2 for weathering. The biggest problem with using metal halide lamps for weathering is that the lamp technology is not standardized. There is no single metal halide lamp due to how the technology works. There are a great many types, similar to the variety of fluorescent lamps manufactured. For weathering, fluorescent UV lamp types have been standardized and can be obtained from multiple sources around the world. No such standardization is yet to occur for metal halide lamps. Furthermore, not enough is known or published about the spectral stability of these lamps at different irradiance levels. Does the spectrum shift at high irradiance versus low? The spectral curves for xenon long arc lamps used for weathering tests experience minimal change between low and high irradiance. How do these lamps age and how long can they maintain the UV spectrum compared to xenon arc lamps? Xenon arc chamber manufacturers have specific lamp life specifications. None of these uncertainties regarding metal halide lamp use in testing is a permanent state. Over time these questions may be answered and specific metal halide lamp spectra standardized. Until then, metal halide lamp use for weathering is still in the realm of experimentation.

276

D URABILITY AND R ELIABILITY

OF

P OLYMERS

11.9 Climatic Tests in the PV Module Qualification Standards Fig. 11.14 shows an example of an environmental chamber designed for testing PV modules. There are three climatic tests in the qualification standards: thermal cycling, humidity freeze, and damp heat. These are labeled MQT 11, 12, and 13, respectively, in IEC 61215-2. MQT 10, UV preconditioning test, is considered separately from these climatic tests earlier in this chapter. These tests specify module temperature rather than chamber air temperature. It is important to note that these are parts of test sequences within the qualification test regime. Sometimes the stresses cause a failure mechanism not detected until later on in the sequence. Therefore, each one is of limited value on its own. The cycling tests stress soldered connections and adhesion between laminated layers, while damp heat detects moisture ingress that results in delamination between layers and corrosion of metallic interconnects. Each is briefly described with an emphasis on chamber characteristics important to it. For brevity, only 61215 sequences are discussed, although the 61730 safety standard series includes similar test sequences with different performance criteria. The damp heat test is the most straightforward, yet widely considered the most challenging. The module is maintained at 85
C with 85% RH in the surrounding air. It is performed in its own sequence,

Figure 11.14 Environmental chamber for testing PV modules. Photo courtesy of Espec.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

with half of the tested modules then undergoing a hail impact test and the other half a mechanical load test. Most labs dedicate chambers for this test for reasons of economy since the test requires no refrigeration. Thermal cycling is performed during two of the test sequences. In one of them, it as an independent test conducted for 200 cycles prior to a final performance measurement. In the other sequence, it is performed for 50 cycles between a UV conditioning test, discussed later, and a humidity freeze test of 10 cycles. The thermal cycling test includes injecting current through the modules, which can detect solder failures in real time during the test. In principle, it is a simple test: module temperature is adjusted between þ85 and 40
C and maintained at each temperature for a minimum dwell time of 10 min. There is flexibility in the ramp rate between conditions. The test should not take longer than 6 h to complete a full cycle from ambient conditions to 40, up to 85
C, and back down to ambient, including the dwell time. However, the transitions cannot be faster than 100
C per hour. One cycle can be completed in as little as approximately 2.8 h rather than the maximum of 6 h allowed. A 200-cycle test can be completed almost 4 weeks faster with the fastest allowed transition rates. As discussed earlier, faster transitions require more heating and cooling power and thus add to the chamber’s cost. In addition, one chamber manufacturer calculates that the faster transition consumes about 50% more power than the slow one and requires 78% more cooling water. The humidity freeze test is conducted after thermal cycling and before tests of the junction box and cable anchorage, although one of the two modules in this sequence skips those tests and goes to final power output measurement tests. This is a combination of the damp heat and thermal cycling tests. The full cycle is 24 h and 10 cycles are required, although by maintaining the maximum ramp rates it is possible to reduce this by around 10%. From ambient conditions, the modules are heated to 85
C, where the RH is maintained at 85% and the conditions are held for 20 h. RH control is turned off and the temperature is decreased to 40
C. The maximum ramp rate is 100
C per hour just as in the thermal cycling test, but is allowed to proceed at double that rate once the modules are at the freezing point. The minimum dwell point at the low temperature is 30 min, and then the cycle is reversed. Many module manufacturers, in the absence of clear alternatives, often report test results from

11: ACCELERATED E NVIRONMENTAL C HAMBERS

AND

longer-duration tests or multiples of the prescribed cycles. Their intent is to show that their modules have improved durability versus older designs or against their competitors. It is an understandable approach. Yet these tests can become very costly, and experts have questioned whether, for example, there is any benefit to the damp heat test beyond the 1000 h in the existing qualification standards. Is it possible that a different approach to improving durability may be worth diverting some of the resources devoted to ever-longer qualification test sequences? Recently the IEC has published a new series of standards that seek to qualify module components and their materials separately from the entire module construction. This is analogous to what car or window manufacturers do. Longer-duration tests on materials are less costly and can deliver better simulations. Once materials have been qualified to withstand sunlight, heat, and moisture, it may be that module qualification only requires testing for mechanical interactions between the components. The existing tests appear effective for this purpose.

11.10 Conclusions This chapter covered the technology of climatic test chambers, from the traditional temperature and humidity chambers to specialized corrosion and weathering devices. The methods for creating the stress factors of heat, moisture, UV light, thermal cycling, and corrosive mists were detailed and related to standardized tests used by the PV module industry. Some of the specific challenges facing the specialized tests were presented in the hopes that users and developers of test methods understand practical limitations and write new methods accordingly.

References [1] Thermotron Industries, Solar PV Testing: A White Paper [Online], July 2010, www. thermotron.com.

T ESTING

OF

PV M ODULES

277

[2] Cincinnati Sub-Zero, Types of Chamber Cooling Systems [Online]. www.cszindustrial.com. [3] Espec North America, Resources: Methods of Humidity Generation [Online]. www.espec. com/na. [4] Cincinnati Sub-Zero, Understanding Humidity in Environmental Test Chambers [Online]. www.cszindustrial.com. [5] Espec North America, Resources: How to Chambers Lower Humidity? Hudsonville, MI: s.n. [6] Wikipedia.org, Hygrometer, [Online]. https://en. wikipedia.org/wiki/Hygro. [7] Q-Lab Corporation, Introduction of Cyclic Corrosion Testing, Westlake, OH: s.n. [8] Laboratory test methods for the simulation of atmospheric corrosion: lessons from the automotive industry, in: S.P. Fowler (Ed.), U.S. Department of Defense Allied Nations Technical Corrosion Conference, SSPC, Birmingham, AL, 2017. Vols. Paper No. 2017e789123. [9] S.S.L. Fowler, Considerations for relative humidity and temperature control in atmospheric corrosion teststandards, Polymers Paint Colour Journal (July 2014). DMG Events. [10] S. Bell, Good Practice Guide No. 124: The Beginner’s Guide to Humidity Measurement, National Physics Lab, Teddington, Middlesex, United Kingdom, 2012.

Further Reading [1] Espec North America, Resources: How Do Test Chambers Get So Cold? [Online], 2016, www. espec.com/na. [2] Arndt, Regan, Pluto, Dr. Ing Robert, Basic Understanding of IEC Standard Testing [Online]. https://www.tuv-sud-america.com/us-en.

12 Outdoor Field Testing: Environmental and System Stress Factors Eric J. Schneller and Kristopher O. Davis Department of Materials Science and Engineering, University of Central Florida, Orlando, Florida, United states

12.1 Introduction to Outdoor Exposure of Photovoltaic Modules The basic operation of a photovoltaic (PV) device is to generate power through the direct conversion of solar radiation. Due to this need for solar radiation, photovoltaic devices are required to perform continuously in an outdoor environment. This operational environment, along with an expected operational lifetime measured in decades, is extraordinarily unique when compared to most other electronic devices. The outdoor environment exposes the device to a wide range of stress factors that place stringent requirements on the photovoltaic device packaging. The encapsulation system works to prevent the harsh outdoor environment from affecting the performance of the active photovoltaic cells without impeding the electrical or optical performance of the internal device. As discussed in previous chapters, the materials used during the encapsulation process are critical in ensuring the long-term reliability and durability of the photovoltaic module. In this chapter, we explore the operational environment experienced by photovoltaic modules. We begin by introducing the stress factors that exist due to the climate in which photovoltaic modules are deployed and the unique stress factors that are introduced from the system design. To understand how photovoltaic modules respond to operational conditions in the field, we discuss the methods for evaluating the module and system performance during and after field deployment. Finally, we provide insights derived from existing work to look at how unique environments impact the reliability and durability of photovoltaic modules.

12.1.1 Importance of Outdoor Exposure The operational lifetime of a photovoltaic module is a critical factor in determining the levelized cost of

electricity (LCOE) for solar energy systems. LCOE is generally expressed in dollars per kilowatt-hours ($/kWh) and is determined by dividing the total lifecycle cost by the total lifecycle energy production [1]. The numerator includes all costs incurred over the lifetime of the system including initial procurement and installation, ongoing operations and maintenance, and decommissioning. The denominator is determined by the cumulative energy production, which is a function of the local climate and system performance over time. Photovoltaic energy systems benefit from minimal operational and maintenance costs as compared to traditional power generation systems because there are no moving parts and no fuel costs. This implies that the overall lifecycle costs are similar, regardless of the operational lifetime of the system. The operational lifetime does, however, have a tremendous impact on the total lifecycle energy production. Simply put, the longer the power generated by a photovoltaic system, the lower the LCOE. Estimates of energy production and resulting LCOE depend heavily on the assumptions used for the module degradation rate and operational lifetime. Real-world operation is the ultimate validation of a photovoltaic module design. In fact, long-term outdoor exposure is the only definitive way to determine the operational lifetime and degradation rate of a module. Unfortunately, this requires a tremendous amount of time that is not realistic for product development cycles. This leads manufacturers to supplement, or even replace, long-term outdoor exposure with indoor accelerated aging to validate module designs. Warranties provided by reputable manufacturers of photovoltaic modules are typically in excess of 20 years. This time frame may or may not be a reasonable assumption for the operational lifetime and will only be determined once systems age beyond this timeframe. In the absence of this long-term performance data, we must rely on

Durability and Reliability of Polymers and Other Materials in Photovoltaic Modules. https://doi.org/10.1016/B978-0-12-811545-9.00012-4 Copyright © 2019 Eric J. Schneller & Kristopher O. Davis. Published by Elsevier Inc. All rights reserved.

279

280

D URABILITY AND R ELIABILITY

OF

P OLYMERS

indoor accelerated aging to infer long-term performance. To make critical assumptions that link indoor accelerated aging to equivalent time frames in the field, it is essential to understand the operational environment of the system and the stress factors that lead to module degradation. Ultimately, real-world operation is the only way to validate these assumptions and is critical for building confidence in the reliability and durability of photovoltaic modules and systems. The first photovoltaic module design validation efforts were carried out by the Jet Propulsion Laboratory in the late 1970’s within the Flat Plate Solar Array Project [2]. In this early work, they identified the key drivers for module degradation to be system voltage, temperature, humidity, ultraviolet (UV) radiation, soiling, salt spray near marine environments, and mechanical stress. Although module technologies have progressed since these times, the stress factors that were identified are still relevant. In the following sections, we explore each of the stress factors that exist during operation and discuss methodologies to evaluate their severity and assess their impact on reliability and durability. It is important to remember that the outdoor environment is a complex combination of multiple stress factors and that every location provides a unique combination.

12.1.2 Environmental Stress Factors Operating temperature is one of the most critical stress factors because it has an accelerating influence on many module degradation mechanisms [3]. Most chemical reactions follow an Arrhenius relationship, where the logarithm of the reaction rate (k) is inversely related to the absolute temperature (T) as shown in Eq. (12.1) [4]. k ¼ AeEa =RT

(12.1)

Ea is the activation energy of the reaction, R is the ideal gas constant, and A is the preexponential factor that describes the relationship between temperature and the reaction rate. As a general rule of thumb, this relationship describes a reaction rate that roughly doubles for every 10 C increase in temperature. Many degradation mechanisms follow this relationship including photothermal degradation of the encapsulant and metallization corrosion [5].

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

In addition to chemical reactions, temperature also plays a role in degradation mechanisms that may be sensitive to differential expansion stresses. As the operating temperature of the module varies, differences in the coefficients of thermal expansion (CTE) of various materials can result in applied mechanical stress. This is commonly referred to as the thermomechanical stress and applies most directly to the cell/solder/ribbon system in which there can be substantially disparate CTEs [6]. In the case of extreme operating temperatures this thermomechanical stress can result in significant force being applied to the cell which may result in cell cracking. Alternatively, the daily cycling of temperatures can result in cyclic mechanical stress that may induce fatigue in the solder joints or within the interconnect ribbon [7,8]. The operating temperature of a module is affected by many variables in real-world systems including the ambient temperature, incident irradiance, wind speed and direction, the module mounting configuration, and the module loading condition (e.g., maximum power, open-circuit, and shortcircuit). The module temperature is commonly used in system performance predictions and several models have been developed to accurately predict module temperature given the system configuration and ambient weather conditions [9e13]. These models are valuable from a reliability perspective because it enables the prediction of expected operating temperatures in a particular location over the lifetime of the system. One common method for calculating the back of the module temperature (Tmod) is to use the model developed by King et al. shown in Eq. (12.2) [10]. Tmod ¼ Tamb þ Geaþbv

(12.2)

Tamb is the ambient temperature, G is the global plane of array irradiance, v is the wind speed, a is a coefficient that accounts for the relationship between the incident radiation and the module temperature, and b is a coefficient that accounts for the cooling effect that wind has on the module temperature. Coefficients a and b are determined empirically for a given module construction, system design, and location. It is also important to note that the back of the module temperature is typically 1e3 C cooler than the cell temperature, which varies based on the properties of the module packaging materials. More thermally insulating packaging (e.g., glasseglass

12: O UTDOOR F IELD T ESTING : E NVIRONMENTAL

AND

construction) will lead to a higher operating temperature and a smaller difference between the rear side and the cell. This temperature model also excludes the effect of radiative cooling at night. The exclusion of this cooling factor is acceptable because degradation primarily occurs at higher operating temperatures. Humidity is another key stress factor that plays a crucial role in several module degradation mechanisms. Humidity refers to the amount of water vapor in the ambient air and can be expressed in either absolute or relative terms. Absolute humidity represents the moisture content in air measured in grams per cubic meter. More common, however, is the use of relative humidity measured as a percent that quantifies the current absolute humidity relative to the maximum absolute humidity possible at a given temperature. For the ambient relative humidity to affect the internal components of the module, moisture must penetrate through the external packaging of the module. Some solar cell technologies, such as thin-film CdTe and CIGS, are extremely sensitive to moisture so the packaging is designed to be hermetically sealed using glasseglass encapsulation with an edge seal [14]. For most silicon technologies moisture is not as detrimental. In this case, it is common to use a polymer based backsheet that allows moisture to permeate within the module. The internal moisture concentration can be determined using a diffusion model, where the ambient relative humidity creates a concentration gradient that drives moisture into the module. The key parameters that dictate the level of moisture within the module include the absolute humidity at the surface of the module, water vapor transmission rate of the backsheet, the water-solubility of the encapsulant, the physical thickness of the packaging materials, and the spacing between cells. There is no direct relationship between the ambient relative humidity and the internal moisture content due to the variation of the encapsulant properties with temperature and the time-delay caused by the relatively low diffusion rates within the encapsulant. Finite element methods (FEM) must be employed to model the moisture ingress within a module [15e17]. The results from such a model are shown in Fig. 12.1. The moisture concentration within the encapsulant varies as a function of time and location within the module. The moisture concentration behind the cells equilibrates with the ambient relative humidity in only a

S YSTEM S TRESS FACTORS

281

Figure 12.1 Results from a FEM simulation showing the moisture concentration within the encapsulant in Freiberg, Germany at two locations within the module [15].

matter of hours, while thousands of hours are required for equilibrium to occur on top of the cells. This behavior results in moisture concentrations that respond to daily relative humidity cycles behind the cell and to seasonal fluctuations in the average relative humidity on the top side of the cells. Humidity, or water concentration, acts as an accelerating factor for degradation. The rate equation shown in Eq. (12.1) can be extended to include a generalized term for internal relative humidity (RH) with a variable reaction order (n) as shown in Eq. (12.3) [17]. k ¼ AeEa =RT *RH n

(12.3)

A study has shown that for hydrolysis of a polymer encapsulant, the reaction exhibits second order kinetics with respect to water concentration [18]. Moisture also plays a key role in metallization corrosion, which occurs through a two-step process. A combination of heat and humidity results in the formation of acetic acid for EVA encapsulants. This highly corrosive degradation product causes a phase segregation within the screen printed metallization increasing series resistance within the cell [19,20]. A breathable module package scheme that allows acetic acid to diffuse out of the module may help to reduce the rate of this type of corrosion. Another critical stress factor is ultraviolet (UV) radiation. UV is known to cause degradation in polymeric materials used for module encapsulation. UV degradation, or photodegradation, typically

282

D URABILITY AND R ELIABILITY

OF

P OLYMERS

causes discoloration and embrittlement of the polymeric encapsulant and backsheet [21,22]. Although this photodegradation does not directly impact the performance of the cell, it can result in a loss of optical transparency reducing the number of photons that reach the device. Embrittlement, or other mechanical degradation of the polymers, can act to accelerate moisture driven degradation pathways within the module. UV driven degradation is typically accelerated with increasing temperatures. The most severe discoloration has been observed in climates where high doses of both UVA (315e400 nm) and UVB (280e315 nm) occur along with high module operating temperatures. This would be typical in a hot and dry, desert-like climate. Finally, there are several environmental factors that can result in the application of mechanical load to a module, the most common being snow and wind loading. Typical module packaging schemes include a thick front glass sheet and a thin polymeric backsheet. This asymmetrical structure results in tensile stress being applied to the cell with a front side mechanical load and compressive stress applied to the cells with a rear side mechanical load. When cells experience tensile stress they become susceptible to cracking or fracture. As snow accumulates on a module surface, a static frontside load is applied to the module. If the magnitude of the load exceeds some critical value, cell fracture will occur, and these cracks will remain as a permanent defect in the module even after the snow has melted and the load is effectively removed. Wind loading, on the other hand, is a much more dynamic stress factor. The direction of the wind determines whether the load is applied from either the front or rear side, and the magnitude of the applied load can change rapidly with wind gusts. This type of stress is simulated better by a cyclic, or dynamic, mechanical load. The magnitude of the applied mechanical load from wind may not be as severe as snow, but the cyclic nature can exacerbate existing cracks and lead to fatigue in solder joints or interconnect ribbons. It is important to note that for alternative module packaging schemes, specifically for glasseglass encapsulation, the cell can exist in the neutral axis. From a mechanical point of view, this means that under any applied load the cell does not experience either compressive or tensile stress. This type of packaging could effectively eliminate the potential for external loads to generate new cell cracks.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

12.1.3 System Specific Stress Factors The design of a photovoltaic system plays an important role when assessing the potential stress factors that influence module degradation. It is useful to consider the region surrounding the module as its own unique microclimate, where the local conditions may vary significantly from nearby ambient conditions. A good example of how system design impacts this microclimate is the effect of the module mounting structure. In an open-rack configuration, typically used for ground mounted utility scale systems, the back of the module is exposed to wind which can help reduce the operating temperature of the module. In contrast, a roof mounted system commonly used for residential PV systems can result in significantly higher operating temperatures in the same climate. The roof acts as both a physical barrier to wind as well as a thermal insulator that traps hot air between the roof and the module. Simulated back-of-the-module temperature data are illustrated in Fig. 12.2 for both an open rack (i.e., ground mount) and an insulated back (i.e., roof mount) configuration. During periods of peak solar irradiance, the difference in temperatures between the two scenarios can be greater than 20 C. This temperature difference can have a tremendous impact on the rate at which degradation occurs.

Figure 12.2 Simulated back-of-the-module temperature for a roof mounted system and an open rack system in Miami, Florida using Eq. (12.2) and typical meteorological year data [23]. Coefficients used for estimating back-of-the-module temperature were taken from King et al. [10].

12: O UTDOOR F IELD T ESTING : E NVIRONMENTAL

AND

Mounting configuration can also have a tremendous impact on the mechanical stress applied to a module due to snow or wind loads. The optimal mechanical support system would be a full perimeter support for the module. This ideal mounting configuration is often not practical from a cost perspective, so modules are typically supported at two points along each long edge of the module frame. The placement of these mounting locations can impact the displacement, and resulting stress, observed for a given applied pressure. For example, as the distance between the support points increases, the displacement in the center of the module for a given snow load will increase. Another example would be to consider a single axis tracking system where the module is supported only at the center of the module. In this case, wind or other loads can cause larger displacements at each end of the module. In each scenario the design of the mounting system has a direct impact on the stress distribution within the module for a given external mechanical load. Another unique stress induced by the system design is the system voltage. The serial connection of modules into strings results in maximum system voltages that are typically 600 V for residential systems and could be several thousand volts for utility scale systems. This system voltage, specifically for modules at the end of a string, can result in a large potential difference between the cell circuitry and the module frame that is grounded. This potential difference has been shown to cause severe performance loss due to the migration of ions within the encapsulation and is referred to as potential induced degradation (PID) [24,25]. The system design has a tremendous impact on the magnitude of the system voltage stress that may be applied to the module. These design related factors include the number of modules connected in series, the clamping or mounting configuration, the type of inverter, and whether the positive or negative end of the string is tied to ground. The presence of moisture on the module surface also has an impact on the magnitude of the system voltage stress. A thin layer of water on the front surface caused by dew, rain, or high relative humidity acts as a conductor that affects the electric field distribution within the encapsulant. Since this degradation mechanism is driven by the migration of ions, any factor that impacts the magnitude or direction of the electric field within the module packaging will have an impact on the rate at which degradation occurs.

S YSTEM S TRESS FACTORS

283

12.1.4 Impact of Local Climate The local climate is the ultimate factor that determines the magnitude of each stress factor experienced by a photovoltaic module. Each climate has a unique combination of weather that generates a specific combination of stressors for a module. Climate zones are commonly classified using the Ko¨ppen-Geiger system [26]. This classification system, originally constructed from the presence of unique vegetation groups, identified five general classifications including tropical, arid, temperate, continental, and polar. These groups are then subdivided based on precipitation and seasonal temperature ranges. This classification provides a good basis for evaluating the stress factors that a typical photovoltaic system may experience. Tropical environments are defined by a minimum temperature and experience both extreme heat and humidity. Arid, or desert, environments are defined by the amount of precipitation and are considered hot and dry. Warm temperate climates experience moderate temperatures and levels of precipitation. Continental climates are those that experience substantial amounts of precipitation in the form of snow in the winter and can experience moderate temperatures in the summer months. For simplicity we will consider four general climate types that represent most locations around the world. These include hot and humid (tropical), hot and dry (arid), cool and dry (continental), and cool and humid (temperate). The four locations within the United States are shown in Fig. 12.3, including Miami (hot and humid), Phoenix (hot and dry), Seattle (cool and humid), and Denver (cool and dry). For each location, typical meteorological year data were analyzed to extract the number of hours a module in an openrack configuration would experience in terms of the ambient relative humidity and back-of-the-module temperature using coefficients from King et al. [10] and Eq. (12.2). The data clearly show how the operating conditions can vary substantially from one location to another. Phoenix, although relatively dry, experiences extremely high operating temperatures in comparison to the other locations. Alternatively, Seattle and Miami experience extremely high humidity on a regular basis. Denver exhibits a much more dynamic range of conditions that range from cool and humid to hot and dry. These external variables will have a dramatic influence on the progression of specific degradation mechanisms.

284

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Figure 12.3 Contours representing the number of hours a module experiences a particular combination of temperature and ambient relative humidity. Module temperatures were calculated using TMY data for each location and an open-rack assumption in Eq. (12.2). Histograms were determined using a bin size of 5 C for module temperature and 5% for relative humidity.

For example, corrosion or other moisture sensitive degradation pathways may develop in a tropical environment such as Miami but not in an arid environment such as Phoenix. Thermally activated mechanisms, such as photothermal degradation of the encapsulant, may be more common in an arid environment. Beyond temperature and humidity there are additional factors that may influence module reliability and durability in a given location. Although all regions in the world experience wind, not all locations will experience snow. For regions known to experience substantial snow loads, a more mechanically robust module (e.g., thicker frame) or a more mechanically secure mounting system may be required. Marine environments also pose a unique design challenge due to the corrosive nature of sea salt. In these environments, elements of the module

that may be susceptible to corrosion, including the glassanti-reflective coating, the module frame, and the module connectors, must be carefully selected to prevent performance degradation. Another factor that is driven by the local climate is soiling [27e29]. Soiling is a generic term that refers to dirt accumulation on the module surface. The specific chemical composition and physical properties of the particles that account for soiling are heavily dependent on the local environment. In the case of soiling, the microclimate is critically important. The presence of nearby construction activities or agricultural facilities will have a tremendous impact on the type of soiling that is present and the magnitude. The amount of soiling is dependent on the type of dust particles, the physical properties of the module surface, the tilt angle of the module, and the frequency and magnitude of precipitation.

12: O UTDOOR F IELD T ESTING : E NVIRONMENTAL

AND

A careful consideration of these influences prior to site selection and system design may help to mitigate the effects from soiling. Currently, modules are designed to pass a single qualification sequence of accelerated environmental testing that does not consider the end use environment. As discussed in this section, each end use environment presents a unique combination of environmental stress factors. A module that performs well in one climate may not be suitable for deployment in another climate. The module and system design requirements for a rooftop system in an arid climate may be substantially different from a utility scale system in a temperate climate. As the market for photovoltaic energy grows, there may be opportunities for module manufacturers to design modules that are tailored for a specific application environment. This would allow manufacturers to select specific materials and fabrication processes to optimize cost and module durability for specific use cases and operational environments.

12.2 Outdoor Testing To evaluate the reliability and durability of modules in the field, the performance of modules is monitored outdoors over relatively long periods of time, compared to that of accelerated aging. This section provides an overview of the physical infrastructure, test methods, and analysis techniques commonly used, and it concludes with some best practices for ensuring that high quality data are collected throughout a long-term performance monitoring experiment.

12.2.1 Equipment and Infrastructure There are a number of ways to configure a PV system to investigate the long-term performance and durability of modules in an outdoor setting. In the design of any experiment, the following should be considered: (1) structural integration of the PV modules; (2) granularity of the monitoring; (3) the electrical load and IeV sensors used; (4) meteorological station; and (5) data acquisition system. Some of the most common options for each are shown in Table 12.1. Additional components and considerations may need to be taken into account, depending on the experiment. For example, if bifacial PV modules are being considered, the albedo of the

S YSTEM S TRESS FACTORS

285

ground cover underneath the modules is another important consideration [30]. If potential induced degradation is being studied, then a voltage source may be applied to the module frame [31]. As explained previously, the structural integration of the system, or how the modules are mounted to a structure, dictates the microclimate the modules will experience. This affects the operating temperature of the cells and modules drastically, as well as the mechanical stress applied to the modules. For the purposes of evaluating how outdoor exposure influences module reliability and durability, the experiments should cover the various methods in which the modules might be installed. From a practical perspective, fixed, ground-mounted systems with an open rack are the most convenient, since the front and rear sides of the modules are both easily accessible. Experiments for rooftop mounted systems can be simplified and made a little safer using faux rooftops installed closer to the ground height. The granularity with which the PV modules are monitored simply refers to whether the IeV characteristics of the module are individually monitored or the modules are connected in series, and then the whole series-connected string is monitored. There are many benefits to monitoring individual modules. When modules are connected in series, the resulting data reflect the performance of the array as an ensemble, and any variance from module-to-module is lost. Additionally, the module producing the lowest current in a series string limits the performance of the entire string (i.e., mismatch loss [32e34]), thereby further adding unwanted complexity when looking to interpret data when an underperforming module is present. When all of the modules are performing well, this is not a critical issue. However, when studying the role of external climate and system stressors and something goes wrong, a higher level of granularity can be very valuable in diagnosing the problem. With many PV systems installed in the field, particularly for large utility-scale projects, series strings are also connected in parallel in combiner boxes. This aggregates the performance of an even larger group of modules, and is therefore not common in scientific experiments focused on module performance and durability. A simple illustration of these different options is shown in Fig. 12.4. When choosing the appropriate electrical load and IeV sensors to monitor the performance of the PV modules, several options exist. Load resistors with

286

D URABILITY AND R ELIABILITY

OF

P OLYMERS

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

Table 12.1 Common PV System Designs and Components Used for Monitoring the Long-Term Performance and Durability of PV Modules  Ground-mounted with a fixed open rack

Structural integration

 Ground-mounted with a tracking system  Rooftop, or faux rooftop, mounted with the modules installed parallel to the roof, with or without an air gap inbetween  Monitor every module separately

PV monitoring granularity

 Connect modules in series and monitor the series-connected strings  Connect modules in series, then combine the series strings in parallel and monitor the series-parallel arrays Electrical load and sensors

 Load resistor with fixed resistance and IeV sensors  Maximum power point tracker with IeV sensors  A curve tracing system with IeV sensors  Grid-connected inverter: string inverter, microinverter, inverter with a DC optimizer, and inverter with a DC string monitoring system

Meteorological station

Any or all of the following sensors:  Irradiance sensor (e.g., reference cell, pyranometer)  Spectroradiometer  Ambient temperature sensor  Module temperature sensor  Anemometer  Humidity sensor  Rain gauge  Datalogger

Data acquisition system

 Network interface  External server or cloud service

PV modules

Individual module monitoring

Electrical load

+

PV modules

Electrical load +

PV modules

Electrical load +

Seriesparallel connected module monitoring

Seriesconnected string monitoring

+

+

-

-

Figure 12.4 Different levels of granularity in monitoring the PV module IeV characteristics.

12: O UTDOOR F IELD T ESTING : E NVIRONMENTAL

AND

S YSTEM S TRESS FACTORS

fixed resistance, coupled with IeV sensors, are the simplest and most economical solution, but have many drawbacks. Often, the resistance value selected is meant to load the module near maximum power point at some predetermined irradiance level (e.g., one sun condition), and when the condition is met, the resulting current and voltage are roughly equivalent to the IMP and VMP of the module or string. However, because irradiance changes throughout the day, the operation point fluctuates and the actual current and voltage therefore fluctuate. This is illustrated in Fig. 12.5, along with examples of the full IeV curve and VMP at 0.1, 0.5, and 1 sun conditions, corresponding to 100, 500, and 1000 W/m2. Another solution is to use a maximum power point tracking system with IeV sensors [35]. This enables IMP, VMP, and PMP to be monitored under varied irradiance and temperature conditions, unlike the fixed impedance load resistor option, as illustrated in Fig. 12.5. Much more information can be gleaned if the entire IeV curve of the module or string is monitored, rather than just one point at a fixed impedance. Curve tracing systems with IeV sensors

287

do just that. They are programmed to vary the impedance seen by the module or string from a shortcircuit to open-circuit condition (not necessarily in that order) and collect the entire curve. This allows for well-known module performance parameters like ISC, VOC, IMP, VMP, and fill factor to be collected. With additional data processing (e.g., curve fitting), additional parameters like RS and RSH can also be extracted. The solutions described previously are commonly used at research institutions that study the performance of PV modules in the field. One downside common to them is that none of the energy produced is actually put to use in any useful (i.e., economic) manner. An alternative is to use grid-connected inverters where the energy produced by the modules can be readily used by electrical loads on-site or delivered to the grid. String inverters and microinverters are both good options, with the primary difference being the granularity at which the PV module is monitored, with microinverters being able to monitor individual modules. DC optimizers or DC string monitoring solutions can also be used to

1 sun I-V curve V

ISC at 1 sun

VMP

Current

Slope R-1

0.5 suns I-V curve ISC at 0.5 suns

ISC at 0.1 suns

VMP V < VMP

0.1 suns I-V curve

VMP

V 20% cell efficiencies for PERC cells from monocast bricks [9]. If this progress continues, monocast wafers grown using cost-effective brick growth methods may well become a common technology in c-Si photovoltaics.

14.3.1 Silicon Wafers

14.3.2 Solar Cells

Typically most c-Si PV modules utilize solar cells fabricated using multicrystalline silicon wafers, due to the increased cost of monocrystalline silicon wafers, even though monowafers produce higher efficiency solar cells than do multicrystalline silicon wafers. In the past few years industrial capacity has been introduced to grow monocrystalline boules for wafer production, and this has appeared initially as PERC solar cells made on monocrystalline wafers. As PERC develops, it is being adapted to multicrystalline wafers also. Cost-effective silicon crystal growth for mono and multicrystalline wafers is a major focus of silicon crystal growth companies. Recently, a comparison of material quality of multicrystalline silicon wafers, across many manufacturers, using convolutional neural networks, has shown the large variability and the future room for improvement in this area [6,7]. Another exciting area in crystal growth is the recent progress in “mono-cast” silicon growth, in which substantially monocrystalline

In PV cell materials and architectures, we will continue to see increases in cell conversion efficiency, which, if they come to market in a costeffective implementation, can broadly impact the cost effectiveness of PV power plants in comparison to non-PV electricity generation. To follow the advances in cell conversion efficiency, the National Renewable Energy Laboratory’s (NREL) best research cell efficiency chart allows one to survey across all cell technologies and track the rate of improvement [10]. In addition, further market segmentation, which allows different cell and module technologies to be deployed in different regions and market segments, is vital to PV penetration. For example, thin film PV technologies typically have beneficial temperature coefficients, which can advantage them for hot climates. The technology roadmap for photovoltaics is not based soley on cell materials and cell architectures though. Instead, all the PV module components, the

14: F UTURE T REND

AND

P ERSPECTIVES

PV power plant system components, and even the design and capabilities of electrical grids and societal policies can impact PV penetration and growth.

14.3.3 Interconnect Materials Cell interconnection technologies have advanced little since the early days of the development of crystalline silicon PV modules by the US Dept. of Energy research and block grant program. Tinned copper ribbons, soldered to the front and backside of the PV cells for crystalline silicon modules, have only recently been successfully challenged by novel approaches. These have included attempts for costeffective commercialization of back-side contacted cells, such as metal wrap through cells, which did not reach large-scale utilization because of cost concerns. More recently the focus has been on shingled cells, in which electrically conductive adhesives interconnect the frontside of one cell that is overlapping the back surface edge of the adjacent cell. This shingling is often combined with half-cell architectures, or even sliver cells which can be onesixth of the size of a typical silicon cell wafer. These shingled cell approaches provide the possibility of more cost-effective cell interconnection, even though long-term reliability is still to be demonstrated in the field. Alternatively, traditional ribbon interconnection is also being used with half-cell crystalline silicon PV cells to reduce the resistive losses arising from the high currents in current high powered modules. With this design, a single 72-cell PERC module can produce 400 W of electricity. By configuring the module using half-cells and a series/parallel interconnection, the I2R resistive losses can be reduced. This approach could be extended to sliver cells, so as to reduce module DC current and increase the operating voltage of modules. In the case of building integrated PV modules, another new market segment, novel interconnection strategies are being pursued to address the application demands for novel installation opportunities and appealing visual appearance.

14.3.4 Module Packaging Materials The long lifetime performance and reliability of PV modules arise almost completely from the packaging materials used to encapsulate the solar cells and

331

their interconnects, and protect them from degradation due to exposure to environmental conditions.

14.3.4.1 Frontsheet The sun-facing frontsheet of a PV module is typically made of a low iron glass, even though some flexible PV modules use a polymeric frontsheet (e.g., a fluoropolymer such as ethylene tetrafluoroethylene (ETFE)). Due to the easy breakage of glass under mechanical impact, glass frontsheets are typically either tempered or heat-strengthened. Tempered glass frontsheets are used in glass/backsheet module constructions where the backsheet is polymeric. In these constructions, the glass is typically 3.2 mm thick allowing the glass to be fully tempered. This tempered glass, if broken, will break into small glass pieces of a few millimeter dimensions. In a module with glass used for both the frontsheet and backsheet, sometimes referred to as a double glass module, thinner glass can be used, due to the improved mechanical strength of this “laminated glass” structure. A typical glass front or backsheet is 2.5 mm thick and is heat-strengthened, since it is not possible to temper large glass sheets of this thickness. Advances in frontsheet glass include texturing the inner surface of the glass to improve glass/encapsulant adhesion and reduce reflectance losses from that interface. In addition, on the outside, surface texturing of the glass or deposition of an antireflection coating has been used to improve light coupling and reduce reflectance. But many times these glass frontsheets are subject to increased soiling and power loss increases with soiling. In addition, organic antireflection coatings applied to glass frontsheets have not demonstrated good lifetime performance, since they can degrade under harsh real-world conditions. One of the most aggressive conditions has been the impact of wind-blown sand in desert conditions on coated frontsheets. The increased power output of utility-scale power plants after cleaning has lead to soling research and the development of metrology tools to assess soling as a function of time. These tools provide information on when operation and maintenance (O&M) personnel should clean soiled modules or let cleaning happen by natural rain storms. Automated module cleaning robots are being developed for large-scale systems when the increased power output can help afford these O&M costs.

332

D URABILITY AND R ELIABILITY

OF

P OLYMERS

14.3.4.2 Encapsulant Ethylene vinyl acetate (EVA) has been the standard encapsulant used in c-Si PV modules since their initial development, due to its low cost. But EVA has the undesirable side effect that hydrolysis of EVA leads to production of acetic acid, which corrodes the interconnect and cell metallization, leading to power loss [11]. A variety of polymers have been used in the laminated glass industry for automotive windshields (polyvinyl butyral) and ionoplast ionomeric polymers used in laminated glass windows. In addition, silicones were studied initially for cell encapsulation, but have not achieved market acceptance, due to the long cure times which have a negative impact on factory production rates. These alternative materials have many desirable properties but have not achieved the necessary price point to displace EVA encapsulants in the market. More recently, polyolefin elastomers (POEs) have been studied extensively, and have entered production for PV modules, due to their similarity to EVA encapsulants, while removing the hydrolytic production of acetic acid. POE encapsulants are considered by many as an opportunity to increase the lifetime of PV modules, without excessive cost impact. Their reliability and performance will become more clear after some years of usage in the field.

14.3.4.3 Substrate (Backsheet) The most popular polymeric backsheets used in PV modules are the nontransparent multilayer films of polyvinyl fluoride/polyethylene terephthalate/ polyvinyl fluoride (PVF/PET/PVF) and PVF/PET/ ethylene vinyl acetate (PVF/PET/EVA). Recently Dupont introduced the first transparent Tedlar PVFbased PV module backsheet, specifically for bifacial solar modules, a market segment that would typically use a double glass module construction. The polymeric backsheets have seen many new material introductions over the past 10 years, due to the strong cost-down push in the industry, and their cost position in the PV module bill of materials (BOM). This led to the introduction of more backsheets with only two polymer types, instead of three polymer types such as PET/PET/EVA backsheets in which the air-side layer is not a fluoropolymer. While innovation drives technological advances and increased market growth for PV, there have been

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

cases where backsheet innovation has turned out to be a fiasco for the industry, as arose with the introduction of polyamide 12 backsheets. These polyamide backsheets, in which the air-side layer is polyamide, were able to pass all the required qualification tests during product development and acceptance. But once deployed in the field, polyamide backsheets exhibited extensive macro-scale cracking after only 5 years in the field, long short of the typical 25 year power warranty on current PV modules. It is believed there are currently on the order of 5 GW of PV power plants in the field that will require repair and remediation due to the industry, and our standards are inadequate to identify this issue prior to their being installed in large volumes [12].

14.3.5 Module Architectures Innovations in module architectures include many approaches including split cells, serieseparallel interconnection strategies, shingled cells, the use of electrically conductive adhesives, and many others. Each of these needs to both bring sufficient value to the module and do this in a cost-effective way. Typically, scientists and technologists have focused on increases in various performance metrics, such as cell efficiency, without explicitly considering the additional costs for the mature technology, and this has led to many efforts, such as back-contact cells, that are not economically viable, even if they are technologically demonstrated. In addition, many technology efforts focus on improving instantaneous performance, for example by application of an antireflection (AR) coating to the frontsheet of a module. But the performance over lifetime is not considered. A large number of organic AR coatings have been developed and sold, but these AR coatings have not yet succeeded at scale because the lifetime of the coating is shorter than the module’s lifetime. The rapid acceptance of PERC c-Si cells, and the design flexibility that PERC brings, means that it is relatively easy to produce both monofacial and bifacial PERC cells, without much additional cost. Therefore the ability of a bifacial PERC cell to collect more light, and produce more electricity, is an opportunity that can be exploited by developing new module architectures that take advantage of the PERC cell’s bifacial nature. This can be achieved by either producing bifacial modules to allow light capture

14: F UTURE T REND

AND

P ERSPECTIVES

from the module’s back side or the introduction of white encapsulants behind the cell that serve to reflect light in the module back towards the cell.

14.3.6 PV Systems As PV modules’ lifetime performance increases and their useful lifetimes become longer, other components of the PV power plant are getting renewed attention to their relative lack of progress in lifetime performance. Two areas of special concern are wires and connectors and the DC to AC inverters. Wires and connectors represent one of the larger reliability problems of PV power plants, which represent a significant contributor to O&M expenses. The inverter reliability is such that most power plant owners pencil in periodic inverter replacement on a typically 10-year timeframe. If these two areas were able to develop products with 25-year lifetimes, in a cost-effective manner, then a major ongoing expense in PV power plants could be reduced.

14.4 Data Science Approach and Improved Reliability Studies As the PV industry scales to being a major, multiterawatt contributor to the world’s electricity, we need to dramatically improve our ability to assure reliability of products whose lifetimes are extending to 50 years [13]. The historical way in which reliability studies have been performed, typically in the laboratory, using standards-based accelerated exposure protocols, is incapable of providing technological guidance for products with 25 to 50-year lifetimes. The results of these small lab-based studies tend to be observational, instead of statistically significant. They tend to only observe the initial steps in what are in reality complex, multistep degradation pathways that play out over decades. And many times, either the accelerated test conditions are too aggressive, activating mechanisms that are never observed in the real-world [14] and leading to an increase in manufacturing costs to address a nonexistent problem, or the accelerated exposures are too weak to activate important mechanisms. Lastly these standards-based accelerated tests are incapable of representing the diverse spectrum of real-world climatic zones and exposure conditions, and many times neglect the characteristic cycling that arises in the real-world as day goes to night and as summer goes to winter. It therefore is not surprising that lab-

333

based studies for PV, which can work for consumer products with 2 or 5 year lifetimes, have many time misled the community in what are substantial ways, as seen in the fateful introduction of polyamide backsheets in c-Si PV modules. One advantage for the PV community is that over the next few years, as we approach a terawatt of deployed PV, we have a much more realistic fleet of PV systems to use for our reliability research. Consider that the electricity produced is continuously measured, as a time-series by the plant owner who is selling that electricity to the market. Today, with satellite-based weather and irradiance known on a 3.5 km pixel size, from companies such as SolarGIS, the temporal exposure conditions that each of these systems experiences are then known [15]. In order to continue to improve the lifetime and reliability of commercial PV modules, data need to be shared between manufacturers and commercial power plant owners. This enables the use of engineering epidemiological approaches for reliability, degradation, and lifetime performance [16,17]. This engineering epidemiological approach provides an opportunity to better design materials for stress conditions and to improve accelerated testing of PV modules and materials. This is the robust direction that PV degradation science and research should pursue [18]. And combining this with advances in data science [19] and big data analytics [20] suggests that a reliable path is ahead for PV innovation and research.

14.5 PV Standards Activities for Qualification, O&M, and System Rating Modules will need a climate zone rating schema in order to achieve 50-year lifetimes all over the world. Ko¨ppen Geiger climate zones [21,22] are a good base. Changing the packaging material, cell type, and module architecture to better fit the stress conditions that a module experiences can reduce the price in more temperate climates and increase lifetime in the harsher climates. This requires a different type of PV qualification and system rating. In order to distinguish the durability and reliability of PV modules, it will be very important to develop a rating system. The rating of PV modules depends on the years of service, and it may be classified as: Class I (40 years), Class II (25 years), and Class III (10 years). Such a system will offer potential

334

D URABILITY AND R ELIABILITY

OF

P OLYMERS

advantages, for example, it will help customers select better quality and safety of modules installed in a house or building with lower risk of causing fire and lower insurance rate.

14.6 Module and PV System Recycling, and Repowering PV Power Plants While PV systems provide a more environmentally friendly method of producing electricity for many years, modules will eventually come to the end of their lifetime and need to be replaced in the next 15e20 years [23]. Globally it is estimated that by 2050, w60e70 Mt of modules at the end of life will need to be disposed [24]. Older PV plants/sites can be repowered by replacing the modules with new modules without necessarily needing to replace the whole infrastructure around the PV module (e.g., racking, transmission). However, there is no clear standard on when a site should be decommissioned from a technology standpoint. A difficulty in this is that module sizes have changed over the years so it might not be a one for one replacement, but modules are producing more power per area than previously. The removed modules need to be recycled to recover materials and to reduce the load in the landfill. c-Si PV modules

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

contain large amounts of glass, silicon, various metals, and polymers. The difficulty in recycling includes deconstructing the module into its different material types for recycling and accounting for the changing materials used in encapsulation. While the European Union has implemented policies around recycling [25], the United States does not yet require any recycling of PV modules. As renewable energy is a rapidly growing industry sector, the recycling of PV modules will provide workforce opportunities and will decrease the need to mining new metals.

14.7 Future Perspectives By conducting systematic durability and reliability studies for PV technologies, improved safety and/or performance standards can be recommended and/or established (Fig. 14.2). This requires a systematic approach to looking at real-world power plants (e.g., power data, field surveys, field retrieval) to understand the actual degradation mechanisms occurring in PV sites in multiple different climatic zones. Then accelerated qualification tests need to be performed that better mimic these degradation mechanisms to get accurate lifetimes for PV modules, cells, and materials. Such studies would examine the chemical, optical, mechanical, electrical, and flammability properties

Figure 14.2 Systematic studies for long-term durability and reliability (for final safety), from materials and components to the PV modules/system.

14: F UTURE T REND

AND

P ERSPECTIVES

and functional performance of individual materials and components and the PV modules/system as a whole. Finally, failure modes and/or mechanisms related to PV’s long-term durability and reliability induced by the change of materials, interfacial and module/ system properties, and performance degradation when exposed to weathering, aging, and water/ moisture can then be identified and/or proposed, to help establish a fundamental understanding of materials science and technology involved in the longterm reliability and safety of solar PV panels.

References [1] P. Choudhary, R.K. Srivastava, Sustainability perspectives e a review for solar photovoltaic trends and growth opportunities, Journal of Cleaner Production (2019). [2] International Technology Roadmap for Photovoltaics, tenth ed., 2019. https://itrpv.vdma.org/. [3] EIA - Annual Energy Outlook 2018, 2018. https://www.eia.gov/outlooks/aeo/. [4] International Energy Agency, World Energy Outlook 2018, Organization for Economic Cooperation and Development (OECD)/IEA, Paris, France, 2018. [5] A. Guterres, The Sustainable Development Goals Report 2018, United Nations, Department of Economic and Social Affairs, 2018. https:// www.un.org/development/desa/publications/thesustainable-development-goals-report-2018.html. [6] M. Demant, P. Virtue, A. Kovvali, S.X. Yu, S. Rein, Learning quality rating of as-cut mc-Si wafers via convolutional regression networks, IEEE Journal of Photovoltaics (2019) 1e9, https://doi.org/10.1109/ JPHOTOV.2019.2906036. [7] M. Demant, P. Virtue, A. Kovvali, S.X. Yu, S. Rein, Visualizing material quality and similarity of mc-Si wafers learned by convolutional regression networks, IEEE Journal of Photovoltaics (2019) 1e8, https://doi.org/10.1109/ JPHOTOV.2019.2906037. [8] Kutsukake, Growth of crystalline silicon for solar cells: mono-like method, in: D. Yang (Ed.), Handbook of Photovoltaic Silicon, Springer Berlin Heidelberg, Berlin, Heidelberg, 2018, pp. 1e20, https://doi.org/10.1007/978-3-66252735-1_35-1.

335

[9] E. Fornie´s, B. Ceccaroli, L. Me´ndez, A. Souto, A. Pe´rez Va´zquez, T. Vlasenko, J. Dieguez, Mass production test of solar cells and modules made of 100% UMG silicon. 20.76% record efficiency, Energies 12 (2019) 1495, https://doi.org/ 10.3390/en12081495. [10] Best Research-Cell Efficiency Chart j Photovoltaic Research j NREL, n.d. https://www.nrel. gov/pv/cell-efficiency.html. [11] L.S. Bruckman, N.R. Wheeler, J. Ma, E. Wang, C.K. Wang, I. Chou, J. Sun, R.H. French, Statistical and domain analytics applied to PV module lifetime and degradation science, IEEE Access 1 (2013) 384e403. https://doi.org/10. 1109/ACCESS.2013.2267611. [12] Y. Lyu, J.H. Kim, A. Fairbrother, X. Gu, Degradation and cracking behavior of polyamide-based backsheet subjected to sequential fragmentation test, IEEE Journal of Photovoltaics 8 (2018) 1748e1753. https://doi. org/10.1109/JPHOTOV.2018.2863789. [13] R. Jones-Albertus, D. Feldman, R. Fu, K. Horowitz, M. Woodhouse, Technology advances needed for photovoltaics to achieve widespread grid price parity, Progress in Photovoltaics: Research and Applications 24 (2016) 1272e1283. https://doi.org/10.1002/pip.2755. [14] A. Gok, D.A. Gordon, D.M. Burns, S.P. Fowler, R.H. French, L.S. Bruckman, Reciprocity and spectral effects of the degradation of poly(ethyleneterephthalate) under accelerated weathering exposures, Journal of Applied Polymer Science 9 (2019). https://doi.org/10.1002/app.47589. [15] T.J. Peshek, J.S. Fada, Y. Hu, Y. Xu, M.A. Elsaeiti, E. Schnabel, M. Ko¨hl, R.H. French, Insights into metastability of photovoltaic materials at the mesoscale through massive IeV analytics, Journal of Vacuum Science & Technology B 34 (2016) 050801. https:// doi.org/10.1116/1.4960628. [16] M.P. Murray, L.S. Bruckman, R.H. French, Photodegradation in a stress and response framework: poly(methyl methacrylate) for solar mirrors and lens, JPE 2 (2012) 022004. https:// doi.org/10.1117/1.JPE.2.022004. [17] H.M. Mirletz, K.A. Peterson, I.T. Martin, R.H. French, Degradation of transparent conductive oxides: interfacial engineering and mechanistic insights, Solar Energy Materials and Solar Cells 143 (2015) 529e538. https://doi. org/10.1016/j.solmat.2015.07.030.

336

D URABILITY AND R ELIABILITY

OF

P OLYMERS

[18] R.H. French, R. Podgornik, T.J. Peshek, L.S. Bruckman, Y. Xu, N.R. Wheeler, A. Gok, Y. Hu, M.A. Hossain, D.A. Gordon, P. Zhao, J. Sun, G.-Q. Zhang, Degradation science: mesoscopic evolution and temporal analytics of photovoltaic energy materials, Current Opinion in Solid State & Materials Science 19 (2015) 212e226. https://doi.org/10.1016/j.cossms. 2014.12.008. [19] L.S. Bruckman, Transformative opportunities from data science and big data analytics: applied to photovoltaics, Electrochemical Society Interface 28 (2019) 57e61. https://doi.org/10. 1149/2.F07191if. [20] Y. Hu, V.Y. Gunapati, P. Zhao, D. Gordon, N.R. Wheeler, M.A. Hossain, T.J. Peshek, L.S. Bruckman, G.-Q. Zhang, R.H. French, A nonrelational data warehouse for the analysis of field and laboratory data from multiple heterogeneous photovoltaic test sites, IEEE Journal of Photovoltaics 7 (2017) 230e236. https://doi. org/10.1109/JPHOTOV.2016.2626919.

AND

OTHER M ATERIALS

IN

P HOTOVOLTAIC M ODULES

[21] M. Kottek, J. Grieser, C. Beck, B. Rudolf, F. Rubel, World Map of the Ko¨ppenGeiger climate classification updated, Meteorologische Zeitschrift 15 (3) (2006) 259e263. https://doi.org/10.1127/0941-2948/ 2006/0130. [22] C. Bryant, N.R. Wheeler, F. Rubel, R.H. French, KGC: Koeppen-Geiger Climatic Zones, 2017. https://cran.r-project.org/web/packages/kgc/index.html. [23] F.C.S.M. Padoan, P. Altimari, F. Pagnanelli, Recycling of end of life photovoltaic panels: a chemical prospective on process development, Solar Energy 177 (2019) 746e761. https://doi. org/10.1016/j.solener.2018.12.003. [24] K. Komoto, et al., IEA PVPS Task 12, Report IEA-PVPS T12, 10, 2018. [25] Directive 2012/19/EU of the European Parliament and of the Council on Waste Electrical and Electronic Equipment (WEEE).

Index

‘Note: Page numbers followed by “f” indicate figures, “t” indicates tables.’

A Accelerated durability test, 305 Accelerated life testing (ALT), 226, 230 Accelerated testing, limitations of, 303e305 Acceleration factor (AF), 101e102 Acetic acid permeation rates, 143e144 Acetic acid transmission rate (AATR), 49 ALT. See Accelerated life testing (ALT) Antireflection coating (ARC), 101, 180, 332 degradation, 187, 188f PV cell, 187, 188f Antisoiling coating, 180 ARC. See Antireflection coating (ARC) ARCO Solar, 9 Atomizing spray, 262 Autocatalytic photooxidation mechanism, for polymers, 304f

B Back-contact cells, 128e129 Backsheet degradation, 153 construction and materials, 153e155, 154t, 155f, 155te156t environmental stressors, 162e163 laboratory stress exposures, 161e170 effect of different weathering stress, 164e169 sequential/combined testing, 169e170 standardized weathering procedures, 163e164, 163f rates and modes, 157 service life prediction (SLP), 170 Backsheet degradation in field modules, 157 differences in degradation by backsheet type, 157e159, 158t, 159f

by climate, 159, 160f by mounting configuration, 161, 161f observed field degradation, 157, 157fe158f Back surface field (BSF), 24e25 Balance of system (BOS) components, 297 Biomass energy, 13 Blending polymers, 224 Boron, 97 Bypass diodes, 246e252, 247f, 248t diode evaluation, 247e248 perceived gaps and PVQAT TG4, 248e252 electrostatic discharge susceptibility, 251e252, 252f junction temperature, 249 qualification tests, 250e251, 251f thermal runaway, 249e250, 249fe250f

C Cadmium telluride (CdTe) solar cells, 35 Carrisa Plains PV Power Plant (1984e1991), 9e10 Cascade system, 259e260 C-AST. See Combined accelerated stress test (C-AST) Cast silicone, 6 Cell cracking, solar, 106e108 Cell hotspots, 188e190, 189fe190f Cell interconnection technologies, 331 Cell metalization corrosion, 108e110 Cell microcracks, 187e188 Charlesby-Pinner equation, 29 Chemometric technique, 60 Chilled mirror dewpoint hygrometer, 265e266 Chinese Renewable Energy Law, 13 Chromophores, 51

Climatic test, in PV module qualification standards, 276e277, 276f Coefficients of thermal expansion (CTE), 106e107, 280 Combined accelerated stress test (C-AST), 170 Combiner box, 229e230, 230fe231f Concentrated photovoltaic systems (CPV), 24 Conduction band minimum (CBM), 97e98 Cone calorimetry, 30 Connectors, photovoltaic. See Photovoltaic connectors Contact corrosion, junction boxes, 238e240, 239f, 240t Corrosion cell metalization, 108e110 PV system, 145 Corrosion chambers, 269e270 CPV. See Concentrated photovoltaic systems (CPV) Crystalline PV modules, 177 Crystalline silicon (c-Si) module, 7 wafer-based solar cells, 119 Crystalline silicon PV module field failures, visual, 178e208, 180f backsheet, 193e198 chalking failures, 197 cracks, 194, 194f damage due to lightning, 197e198, 198f delamination and bubble formation, 195e197, 197f moisture ingress from, 197, 198f physical damage, 198, 198f protocols for backsheet qualification, 196t yellowing/browning, 194e195, 195f cable, 206, 206f

337

338

Crystalline silicon PV module field failures, visual (Continued )

cell-to-cell busbars and string-to-string bussing ribbons, 192e193, 193fe194f connectors, 206e207, 207fe209f delamination between encapsulant and glass, 185e186, 186f of encapsulant/PV cells, 185, 185fe186f diode, 204e205, 205f encapsulants, 182e186 yellowing/browning, 183e184, 183t, 184f frame, 199, 200fe201f glass, 178e208, 179t, 180fe183f insulation patch failures, 198, 199f JBox base and lid, 203e204, 205f junction box sets, 199e208 electrical connection inside, 201e202, 201fe204f pottant, 205 power degradation, 208e213, 209t affecting, 209e210, 210f rate, 210e211, 210f rate by climate zones, 211e213, 212fe213f, 214t rate by installation years, 211, 211f PV cells, 186e192 ARC coating degradation, 187, 188f cell hotspots, 188e190, 189fe190f cracks and snail trails, 187e188, 188f light induced degradation, 192 potential induced degradation, 190e191, 191fe192f silver finger oxidation, 187, 187f sealant, 198e199, 199fe200f water ingress damage, 208, 209f Crystalline silicon PV modules, 135 c-Si PV modules, 190e192 packaging in, 153

D Damp heat exposure mini-module PV degradation pathway model under, 61e62, 62f PV degradation pathway model under, 60e61, 61f Damp heat (DH) test, 120, 238e240 Data science approach, 55, 333 DC field sensitivity study failure rates, 217, 219t Degradation. See also specific types of degradation

I NDEX

ARC coating, 187, 188f ethylene vinyl acetate, 138, 138f, 145e146 in fielded PV modules, 52e53, 53f photo-degradation, 137 polyethylene terephthalate, 54e55 fixed and mixed effects regression modeling of, 63e65, 65fe66f multivariate multiple regression modeling of, 65e67, 67fe68f netSEM of, 62e63, 63fe64f parallel factor analysis of fluorescence spectra for, 67e69, 68fe69f polymer, 221, 224e225 potential-induced degradation, 52, 52f Degradation, backsheet. See also Backsheet degradation differences in, by backsheet type, 157e159 effect of light intensity and wavelength on, 164e166, 165f effect of other stressors on, 169 effect of temperature on degradation, 166, 167f effect of water on, 167e169, 168f indicators, 160f modes observed in field modules, 158f rates and modes, backsheet, 157 visible backsheet, 157 Degradation rate analysis, 291 Delamination, PV modules, 146e147 Devices under test (DUT), 258 DH induced degradation (DHID), 120e121 Discoloration, of encapsulant and backsheet polymers, 51 Dry bulb sensors, 264e265 DuPont Photovoltaic Solutions, 28

E ECAs. See Electrically conductive adhesives (ECAs) Elastomers, 224 Electrically conductive adhesives (ECAs) materials and PV modules, 129 in novel cell interconnection strategies, 128e129 performance of PV modules utilizing, 129 shingled cells and subcells interconnected with, 129 wrap-through cells and interconnection, 128e129 Electroluminescence images, of photovoltaic modules, 120, 120f

Electronic sensors, 265 Electrostatic discharge susceptibility (ESD), bypass diodes, 251e252, 252f Encapsulants, 27 chemical aging mechanisms of, 137e139 discoloration of, 142 influence of lamination process on the reliability of, 140e142, 141fe142f material properties in PV module degradationdmaterial interactions, 140e147 physical aging mechanisms of, 139e140, 139f Energy Conversion Devices (ECD), 10e11 Energy Reorganization Act of 1974, 6 Energy Research and Development Agency (ERDA), 6 Environmental chambers, 267e269 Environmental degradation, of thermoplastic polymers, 220e221, 221f Environmental stress factors, 280e282 for PV modules, 177e178, 178f Environmental stressors of backsheet, 162e163 ERDA. See Energy Research and Development Agency (ERDA) Ethylene Propylene Diene Terpolymer (EPDM), 224 Ethylene vinyl acetate (EVA), 7, 14, 27 additives and stabilizers used in, 137t adhesion, 146 degradation, 138, 138f, 145e146 discoloration, 37e39, 38f films, 135 hydrolysis of, 39 hydrolytic degradation of, 52 polymer, 49 PV module encapsulation using, 135e136, 136t yellowing of, 142e143, 144f Young’s modulus of, 146f EVA. See Ethylene vinyl acetate (EVA) External quantum efficiency (EQE), 100e101

F Fick’s law, 30 Field experience, 291 degradation rate analysis, 291 Finite element analysis (FEA), 54e55

I NDEX

Finite element methods (FEM), 281 Fixed and mixed effects regression modeling, of PET degradation, 63e65, 65fe66f Flammability, 30e31 degree of flammability, 30e31 flammability index, 31 Flat-Plate Solar Array (FSA) Block Program, 8, 9t Flat-Plate Solar Array Project, 6e7, 135 encapsulation task of, 7 JPL-FSA project, 8e9 Reliability Physics Task, 7e8 Fluorescent UV weathering chambers, 270e272, 271fe272f Fluoropolymer, 39, 153e154, 155f, 158t, 169t Fourier-transform infrared spectroscopy (FTIR), 157e159, 221 Front-side silver grid corrosion, 120e125 characterization, 122e125, 122fe125f mechanisms, 120e121, 121f FSA. See Flat-Plate Solar Array Functional Service Life, 298, 307

G GHI. See Global Horizontal Irradiance (GHI) Glassebacksheet modules, 142e143 Glasseglass module, 142e143 Global Horizontal Irradiance (GHI), 29 Global warming, 23

H Hotspots, cell, 188e190, 189fe190f Hot Water Extraction Method (HWEM), 39 Humidity chambers, temperature and, 258e261 generation, 259, 261e262 methods of adding, 261 atomizing spray, 262 heated water bath, 262 steam generation (boiler), 261e262 ultrasonic nebulizers, 262 methods of removing, 262e264, 263f Humidity, measuring relative in environmental chambers, 264e266, 264fe265f chilled mirror dewpoint hygrometer, 265e266 electronic sensors, 265 psychrometric wet bulb/dry bulb sensors, 264e265

339

Hydrolysis of ethylene vinyl acetate, 39 of vinyl-acetate monomers, 39

I IEA-PVPS task, 12 Interconnect materials, 331 International Energy Agency (IEA) (1974ePresent), 11e12 Ionomer encapsulant, 185e186

J Jet Propulsion Lab (JPL), 240 Jet Propulsion Laboratory (JPL) Flat-Plate Solar Array Project (1975e1985), 6e9 JPL-FSA project, 8e9 JPL Flat-Plate Solar Array Project, 11 Junction box adhesion, 252e253 adhesive change, 252e253, 253f adhesive creep, 253 Junction boxes (jboxes), 235 history, 235e240, 236f cold impact toughness, 237, 237f contact corrosion, 238e240, 239f, 240t creep and consolidation, 237e238, 237fe238f

K Kirchhoff’s law, 99 Ko¨ppeneeGeiger climatic zones, 47, 55e56

L Levelized cost of electricity (LCOE), 3, 4f, 279 LID. See Light induced degradation (LID) Lifetime and degradation science (L&DS), 48, 55e69, 56f case study fixed and mixed effects regression modeling, 63e65, 65fe66f mini-module PV degradation pathway model, 61e62, 62f multivariate multiple regression modeling, 65e67, 67fe68f netSEM modeling, 62e63, 63fe64f parallel factor analysis of fluorescence spectra, 67e69, 68fe69f PV degradation pathway model, 60e61, 61f statistical data analytics

fixed, random, and mixed effects regression modeling approach, 59 multivariate multiple regression modeling approach, 59e60 netSEM modeling approach, 57e59, 57f parallel factor analysis modeling approach, 60 Light induced degradation (LID), 192, 300 Limited oxygen index (LOI), 30 Long-term sequential test (LST), 32 Long-wave infrared (LWIR) radiation, 100 Low-Cost Silicon Solar Array Project, 6 Low-Cost Solar Array Project, 240

M MAST. See Module Accelerated Sequential Test (MAST) MCC. See Microscale combustion calorimeter (MCC) Metal halide light sources, 274e275, 274fe275f Metallization corrosion, 145e146 Metallization elements, of solar cells, 49f Microscale combustion calorimeter (MCC), 30e31 Mini-module PV degradation pathway model, 61e62, 62f Module Accelerated Sequential Test (MAST), 169e170 Module Performance and Failure Analysis Task, 7e8 Motivators, PV durability and reliability, 329, 330f Multiconductor cables, 226 Multivariate multiple regression (MMR) modeling, 59e60 of PET degradation, 65e67, 67fe68f

N Nebulizers, ultrasonic, 262 netSEM. See Network structural equation model (netSEM) Network structural equation model (netSEM), 48 of PET degradation, 62e63, 63fe64f Nonfluoropolymer, 153e154, 155f, 158t, 169t Novel cell interconnection strategies, 128e129 Nylon 12, 230e231

340

O Ohm’s law, 99 Organic photovoltaics (OPV), 302 OTR. See Oxygen transmission rate (OTR) Outdoor exposure of photovoltaic modules, 279e285 environmental stress factors, 280e282 impact of local climate, 283e285, 284f importance of outdoor exposure, 279e280 system specific stress factors, 282e283, 282f Outdoor testing, 285 equipment and infrastructure, 285e288, 286f, 286t, 287f test methods and analysis techniques, 285e288, 286f, 286t, 287f best practices, 290e291 complementary characterization techniques, 290 data processing and analysis, 289e290 time-series data, 288e289, 289f Oxidation, PV cells silver finger, 187, 187f Oxygen bomb calorimetry, 30 Oxygen transmission rate (OTR), 49

P Parallel factor (PARAFAC) analysis of fluorescence spectra for PET degradation, 67e69, 68fe69f modeling approach, 60 Passivated emitter rear cell (PERC), 24, 119 cell technology, 26f Photo-bleaching effect, 142e143 Photoconversion efficiency, 329 Photo-degradation, 137 Photoluminescence (PL), 100e101 Photo-oxidation, 137 Photovoltaic (PV) backsheet, 153, 162, 164e165, 168e170 degradation rate in, 166 light intensity and wavelength effects on, 165 photodegradation of, 165 temperature effect on degradation of, 166 Photovoltaic cells silver finger oxidation, 187, 187f Photovoltaic components field failures for arcing, 36, 36f cell or interconnect break, 33, 33f

I NDEX

corrosion, 32, 33f defective bypass diodes, 36 delamination, 34, 35f encapsulant discoloration, 33, 34f hot-spot heating, 34e35, 35f inverter, 36e37 junction box failures, 33e34, 34f mechanical damage/glass breakage, 35, 36f PV connector, 37 solder bond failures, 33 polymer materials related problems creep, 39 embrittlement, 39 EVA discoloration, 37e39, 38f interfacial adhesion, 39e40, 39f sealant degradation, 40 thermal expansion, 40 Photovoltaic connectors, 37, 220e223, 221fe222f, 240e246, 241f, 241t, 242fe243f factory vs. field assembly, 222e223, 223f, 223t, 224f field crimping issues, 244 ground faults, 245e246, 245fe246f incomplete connector engagement and fretting corrosion, 243e244, 243f new failure modesdwithdrawal force, 222 rodent and insect damage, 242f, 244 Photovoltaic degradation pathway model, 60e61, 61f Photovoltaic devices, 97e98 cell metalization corrosion, 108e110 mechanistic investigations, 108e109 possible mitigation schemes, 109e110 characterization techniques, 98e101 current vs. voltage characteristics, 99e100, 99f imaging and spectral techniques, 100e101 light-induced degradation, 104e106 mechanistic investigations, 104e105, 105f possible mitigation schemes, 105e106 potential-induced degradation, 101e104 mechanistic investigations, 102e103, 103f possible mitigation schemes, 103e104 semiconductor growth and doping, 97e98 solar cell cracking, 106e108

mechanistic investigations, 107 possible mitigation schemes, 107e108 Photovoltaic durability and reliability motivators, 329, 330f Photovoltaic energy, 3, 5, 279 terawatt, 15e16 Photovoltaic industry, 297 durability vs. reliability, 301e302 improved reliability stud, 333 standardized testing, drawbacks, 53e55 Photovoltaic installation to investors, cost of, 297 Photovoltaic materials accelerated life tests, 299 comparative tests, 299 innovations in, 329e333 and module failure modes, 42 observed failure modes of, 37t performance and service life prediction, 40e42 correlation between outdoor performance, 41, 41f correlation between the degradation of PV materials, 42 development of improved or new materials, 40 laboratory accelerated testing, 41 PV service life prediction, 42 qualification test methods, 40e41 qualification test, 298e299 Photovoltaic megawatt-scale power station, 177 Photovoltaic modules architectures, 332e333 degradation, 14e15 durability and reliability study accelerated testing conditions, 32, 32t experimental setup, 31e32 moisture ingress process, 30 physical aging process, 29e30 polymers/materials flammability, 30e31 of PV materials, 32e40 solar irradiance, 29 thermal process, 29 UV crosslinking and degradation reaction, 29 durability assessment, challenges, 302 effects in, 28 encapsulation system of, 5e6 encapsulation, using ethylene vinyl acetate, 135e136, 136t innovations in, 329e333 installations, 3, 4f

I NDEX

interconnection, degradation, and failures, 119e120 layered construction of, 154f long-term reliability and safety, 334f material functional requirements, 26e28, 26t encapsulants, 27 substrate (backsheet), 28 superstrate (front sheet), 27 packaging materials, 331e332 encapsulant, 332 frontsheet, 331 substrate (backsheet), 332 polymeric materials in, 299, 302e303 qualification testing, 299 reliability vs. durability, 28e29, 28f standards for polymeric materials used in, 43t structure and materials for, 24e25, 25f, 25t testing, 257, 260, 263e264, 271e272 21st century PV module cost reduction, 12e14 Photovoltaic Performance Programme (PVPS), 11e12 Photovoltaic, pn-junctions and, 98, 99f Photovoltaic polymers, 30, 31f, 39 Photovoltaic power plant, 217, 218f, 329 operation and maintenance costs of, 297 repowering, 334 Photovoltaic Power Systems Programme (PVPS) (1993ePresent), 11e12 Photovoltaic Quality Assurance Task Force (PVQAT), 248e249 Photovoltaic service life prediction, 42 Photovoltaics for Utility Scale Applications (PVUSA) Project (1986e1996), 10e11 Photovoltaic shipments, in peak power, 23, 23f Photovoltaic standards activities, 333e334 Photovoltaic system, 333 architectures, innovations in, 329e333 Bathtub curve in reliability engineering of, 47e48, 48f corrosion, 145 long-term reliability and safety, 334f recycling, 334 Photovoltaic technology development campaigns and initiatives, 6e14 insertion, and growth, 5e6 Photovoltaic wire (above ground), 223e227, 225f new failure modes

341

approach for determining bend radius, 226e227, 227t cable subjected to flexing, 226 roundness of cables (filled vs. unfilled), 226 slip force, 225e226, 226f UV robustness, 225 Photovoltaic wire (below ground), 227e228, 228f new failure modes, termite resistant cables, 228 Physical aging, 29e30, 139e140, 139f polymer, 29e30 Physics of failure (PoF) of semiconductors, 302 PID. See Potential-induced degradation (PID) pn-junctions and photovoltaics, 98, 99f POE. See Polyolefin elastomer encapsulants (POE) POEs. See Polyolefin elastomers (POEs) Polyamide (Nylons), 220e221 Polycarbonate (PC) polymers, 54e55 Polyethylene terephthalate (PET) film, 153e154 polymer, 62 Polyethylene terephthalate degradation, 54e55 fixed and mixed effects regression modeling of, 63e65, 65fe66f multivariate multiple regression modeling of, 65e67, 67fe68f netSEM of, 62e63, 63fe64f parallel factor analysis of fluorescence spectra for, 67e69, 68fe69f Polymeric materials, 23e24 in PV modules, 299, 302e303 water vapor transmission rate of, 316 Polymer-related failures in PV modules, 49e53, 49fe50f, 53f delamination and mechanical failures, 50e51 discoloration, 51 potential-induced degradation (PID), 52, 52f Polymers autocatalytic photooxidation mechanism for, 304f blending, 224 composites, 224 degradation, 221, 224e225 flammability of, 30, 31f Polyolefin elastomer encapsulants (POE), 136 films, 136

Polyolefin elastomers (POEs), 27, 40 Polyvinyl butyral (PVB) polymer technology, 27 Poly vinyl chloride (PVC), 220e221 Potential-induced degradation (PID), 52, 52f effect, 143e145 Power degradation, PV module visual field failures, 208e213, 209t affecting, 209e210, 210f rate, 210e211, 210f rate by climate zones, 211e213, 212fe213f, 214t rate by installation years, 211, 211f Predicting functional service life, 305e322 Predictive models, 53e54, 63, 65e66, 65fe66f Psychrometric wet bulb, 264e265 p-type cells, 5 Pulsed thermography, 100 PVDF, 231 PVPS. See Photovoltaic Performance Programme (PVPS) PVUSA Project. See Photovoltaics for Utility Scale Applications (PVUSA) Project (1986e1996) PVUSA test conditions (PTC), 10e11

R Radiation solar, 164 ultraviolet, 137, 162 Relative thermal index (RTI), 24 Repowering PV power plants, 334 RH control, temperature and, 266e269 Ribbon fatigue, 127e128, 127f characterization, 128, 128f mechanisms, 127e128 RTI. See Relative thermal index (RTI)

S Salt mist, 269e270, 270f SCC. See Stress corrosion cracking (SCC) Schwarzschild’s law, 314 Sealant degradation, 40 Semiconductors, 98 growth and doping, 97e98 physics of failure of, 302 Sensors electronic, 265 psychrometric wet bulb/dry bulb, 264e265

342

Service Life Prediction (SLP), 305e306, 306t of backsheet, 170 calculate service life estimate, 321, 322f identify degradation modes and pathways, 307, 308f initiate long term in-service and accelerated natural weathering, 318 in-service environment, 308e310, 309f, 309t, 310f model time-to-model time-to-failure as a function, 320e321, 321t practical limitations of, 322 quantify effects of weathering stresses, 310e311 quantitatively define failure (end of functional life), 307, 308f reciprocity, 314e315, 315f test to failure, 318e320, 319f validate SLP model, 321e322 weathering stressdcombined, 317e318 weathering stressdlight, 311e312, 311fe312f weathering stressdmoisture, 315e317, 317f weathering stressdtemperature, 312e314, 313fe314f Shingled cells and subcells interconnected with ECAs, 129 Shockley-Read-Hall (SRH) mechanism, 99e100 Silicon, 97e98 solar cells, 5, 27, 97, 119e120, 119f wafers, 330 Silicone, cast, 6 Slip force, 225e226, 226f Snail tracks, 145 Solar arrays, 5 Solar cells, 97, 104, 330e331 cracking, 106e108 crystalline silicon wafer-based, 119 encapsulation, materials, 136t metallization elements of, 49f silicon, 5, 27, 97, 119e120, 119f Solar irradiance, 29 Solar radiation, 164 SPHERE weathering chamber, 274e275 Sputnik, 5 Starfish Prime nuclear test, 5

I NDEX

Steam generation (boiler), 261e262, 261f Stress and response framework, 56e57 Stress corrosion cracking (SCC), 230 Stress factors, environmental, 280e282 Structural equation modeling (SEM), 48 Sunlight Program, 12e13 SunShot, 3, 4f Sustainable development goals (SDG), 15

T

TedlarÒ, 28 Temperature and humidity chambers, 258e261 air circulation, 258e259 heating and cooling, 259e260 humidity control, 260e261, 260f temperature control, 259 and RH control, 266e269 environmental chambers, 267e269 transitions between test conditions, 266e269, 266f Terawatt PV energy, 15e16 Terrestrial solar power, 5 Thermal cycling, 162e163, 276 Thermal runaway, bypass diodes, 249e250, 249fe250f Thermography, 100 pulsed, 100 Thermogravimetric analysis (TGA), 61f Thermoplastic encapsulants, 136 Thermoplastic polymers, environmental degradation of, 220e221, 221f Thermoplastic polyolefin encapsulants (TPO), 136 Thermoplastic silicon elastomer (TPSE), 147 ThermoPlastic Urethane (TPU), 220e221 Thermo-putative degradation, 125e127 characterization, 126e127 mechanisms, 126, 126f Thermosets, 224 Time-of-wetness (TOW), 316 TPO. See Thermoplastic polyolefin encapsulants (TPO) TPSE. See Thermoplastic silicon elastomer (TPSE)

Transparent conducting oxide (TCO), 316e317

U UCET. See Universal Chemical Exposure Test (UCET) UL94 test, 30 Ultrasonic nebulizers, 262 Ultraviolet (UV) absorber, 143 light, 162 radiation, 162 stabilizer, 63 radiation, 137 UL 1703 Water Spray Test, 238e240 Unit under test (UUT), 303 Universal Chemical Exposure Test (UCET), 230e231 USSR, 5

V Valence band maximum (VBM), 97e98 Visible backsheet degradation, 157

W Water vapor transmission rate (WVTR), 30, 49, 143e145, 145t of polymeric materials, 316 Weathering chambers, 270e275 fluorescent UV, 270e272, 271fe272f metal halide light sources, 274e275, 274fe275f xenon arc chambers, 272e274, 273fe274f WilliamseeWatts exponential function, 29e30 Wire harness, 217, 228 Wire management, 217 devices, 217, 230e231, 231f Wire splices and in-line fuse holders, 228e229 Wrap-through cells and interconnection, 128e129 WVTR. See Water vapor transmission rate (WVTR)

X Xenon arc chambers, 272e274, 273fe274f Xenon arc lamps, 272e273